Research On Combustion Monitoring Modeling In A Circulating Fluidized Bed Boiler

In recent decade years, the circulation fluidized bed (CFB) boiler has been one of the most popular clean coal combustion technologies, which catch more attention of scientists under the requirements of energy conservation and emissions reduction. Because of its disadvantages of nonlinearity, time varying, coupling and distribution parameter, it’s very difficult to establish accurate models. In this paper, models about desulfurization and combustion are established respectively. The main work and research results are displayed to be five points below.1) By the deep research of desulfurization process of an domestic CFB boiler, the existence form of sulfur, precipitation process of sulfur dioxide, existence state of limestone and influence factors of desulfurization efficiency are discussed in this work and decide to divide the desulfurization process to calcinations and sulphation to discuss respectively. The establishment of monitor models about limestone existence state is meaningful to understand the desulfurization conditions and predict the SO2 emissions.2) Considering that the limestone particles can’t be fully used and also be affected by the combustion process, calcining utilization of active CaCO3 ηCoCo2 and sulphating utilization of CaO ηCaO are mentioned, and then the compute and prediction methods are given. Based on above work, storage models about active CaCO2 and CaO are set up. Then the soft sensing model for predicting SO2 emissions is established.3) Considering the time delay of limestone control system, a nonlinear proportional-integral-derivative controller is designed based on the established models to accurately control the feed rate of limestone. The simulation results show that the established models is consistent with experimental results and the designed control system can realize the control target. All of above can guide the actual running of limestone feed rate of power plant.4) As combustion process involving combustion science and hydromechanics, it’s harder to describe the combustion by models, an advanced artificial intelligence (Adaptive-tree-structure-based fuzzy inference system) is used to establish the black-box model for boiler efficiency, NO* and SO2 emissions. Then the accuracy of this model is compared with the artificial neural network and support vector machine.5) To further realize the target of energy conservation and emissions reduction, based on the above model, fruit flies optimization algorithm is used to optimize adjustable parameters aim to ensure the higher boiler efficiency and lower air pollutions emissions. The results show that the optimization results are consistent with the prediction. It’s meaningful to the actual combustion and optimization of power plant.