Boiler Combustion Optimization Model And Its Verification

How the coal-fired boiler can to achieve high boiler efficiency and reduced NOx emissions as the most concerned about for the power plants. Through combustion optimization can be very good to improve boiler efficiency and reduce NOx emissions, so the contents of this paper has important practical significance.For the analysis of DaTong No.2Power Plant#7before and after the burn unit of hedge wall around the boiler combustion unit operating conditions to establish Combustion optimization model. Through the analysis of plant operating data, in the divided condition fixed load range,(similar to the coal and the outside ambient temperature) in the case, pulverized coal concentration and flow rate, the secondary air flow and wind burn, less warm water flow and gas block opening the control panel combustion is the main influence factors. It can be adjusted by controlling the appropriate amount to achieve optimized combustion power plant. Plant-level monitoring system allows plants to save vast amounts of historical data, cleaning method by clustering to analyze the historical data, to find the optimal range of operating history,build a rule-based combustion optimization model library and run continuously improve this model to ensure real-time nature. By the corresponding control program to run quickly to the nearest optimal conditions close range, complete closed-loop control by DCS to ensure that as much as possible in the optimization of its operating range, thus completing the power plant optimization objectives.Using support vector machine in the small sample the superiority of the study and set up a corresponding support vector machine burning optimization model, with genetic algorithm to realize optimization of the parameters for the model, through the model prediction of the status of the boiler, and the experiment prove the establishment of the support vector machine model has good prediction ability. Clustering based data mining is a large amount of historical data of power plant directly, there might be some rules of operation in power plant operation to the practical situation, and to the support vector machine and genetic algorithm optimization combustion modeling prediction function on is a complementary, both supplement each other. Through the Visual C++programming realize optimized combustion of the development of software, and in Da Tong two electric7#boiler put into use, realize the goal of multiple objective optimization, and improve the boiler efficiency and reduce NOx emissions.