| With the development of society and economy, the Chinese government keeps promoting the electric power industry reform, which brings out some new changes.The grid needs to enhance its peak load regulation capacity because of the increasing ratio of new energy resources in its energy structure. The annual operation time of fossil fuel power generation units decreased greatly because of the reduction of electricity demand of heavy chemical industry. Thermal power plants struggled to meet the increasingly tight NO emissions limits imposed by the government. As the dominate energy resources in grid, fossil-fuel power plants also have to face some new challenges. For example,(1) The power plants should help enhance the peak load regulation capacity by operating at a variable load mode.(2) The power plants should make full use of their operation time to maximize their profits by improving operation efficiency.(3) The power plants should strictly control the NO emissions, while also maintain a relatively high operation efficiency.To meet these requirements, thermal power plants needs to develop new techniques to improve their operation flexibility. During the variable load operation, the power generation unit should satisfy multiple objectives such as minimizing the load tracking error, improving operation efficiency, reducing NO emissions, etc.In the power plant load tracking system, the furnace combustion process is the most important and promising part to improve operation efficiency and restrict NO emissions. However, in the current power plant load tracking control systems, there is a lack of furnace combustion optimal adjustments due to the following factors:(1) There is a lack of control-oriented nonlinear dynamic modeling of boiler-turbine units,which include the detailed dynamics of the furnace combustion adjustments.(2) The power plant load tracking control system has to optimize multiple objectives, and there is a lack of research and application of advanced multi-objective optimization control algorithms in this system.To the aforementioned problems, a boiler-turbine load tracking control system with simultaneously boiler economic low NOx combustion optimization is proposed in this paper.To fulfill the design of this control system, a drum boiler-turbine unit with direct-fired mills is selected as an example in this research. The research is carried out with studies on control-oriented modeling and advanced control algorithms as well as some simulations to validate the efficiency of the designed the control system. The main contents of the paper are listed as follows:Research on modeling-(1) The main dynamics of the boiler system and the interactions of different subsystems were studied. In addition, the chemical reaction mechanisms related to NO formation and deduction and the heat and mass transfer principles of flue gas in the furnace combustion process were carefully analyzed, which helped determine the main manipulated variables for the furnace combustion adjustments. Based on the analysis,two models depicting the dynamic interactions of both the boiler efficiency and NO concentrations with furnace combustion adjustments were proposed with pure mechanisms and grey-box modeling methods, respectively.(2) The information from the in-situ data is fully exploited. A soft measurement method,which calculates the furnace combustion state variables by using the measurable parameters of the boiler water-steam subsystem, was proposed based on the interactions of different subsystems. Additionally,a nonlinear dynamic model of the boiler-turbine unit depicting the furnace combustion adjustments dynamics was also presented.Research on control algorithms-(1) A stable multi-objective economic model predictive control algorithm was proposed for boiler-turbine unit load tracking system. This algorithm adds the economic objectives, related to the power generation process, to the feedback control loop for optimization calculations, whose stability and efficiency was validated by both proofs and simulation studies.(2) A preference-adjustable multi-objective model predictive control algorithm was proposed to solve the problem, related to objective priorities,in most multi-objective optimization algorithms,where the tuning parameters have unclear relationships with the objective priorities and thus are hard to be selected. The proposed control algorithm was proven to stable and efficient by both theoretical analysis and simulation studies.Research on control system design and simulations-Based on the fundamental studies in modeling and control algorithms, a boiler-turbine unit load tracking control system with simultaneously boiler economic low NO combustion optimization was constructed. The simulation studies indicate that the designed control system performed well in several aspects, such as improving unit loading tracking ability,increasing unit operation efficiency, reducing NO emissions, etc. |
PDF Full Text Request