Research Of Rough Set Theory On Controlling NO_x Emissions In Coal Fired Utility

As a result of the increasing seriousness of environmental pollution, the standards of NO_x emissions are more stringent than ever before. Although there are various ways to lower the NO_x emission, it without doubts that reduced NO_x formation in furnace can be more economical.In this article, studies of low NO_x combustion optimization are firstly discussed and summarized, and it points out that the current methods of how to control the emission of NO_x mainly includes the combustion optimization and the flue gas treatment. Compared with other artificial intelligence algorithm, such as neural network algorithm, Rough set data mining algorithm are much easier than the Neural network’s, and optimal range of parameters are obtained, which can reduce NO_x generation at the combustion phase.Then, an improved algorithm of attribute reduction based on discernibility matrix and value reduction algorithm based on kernel value introduced in this paper. Improved attribute reduction algorithm does not require complex logic operation, easy to use programming language to implement, and less memory used to store distinguish function, the efficiency of the program improved greatly. Due to the Value reduction algorithm, regulation rules are further streamlined. At the same time, in order to show the mining process and the reduction result more intuitive, visual studio are used in this paper to programmed the attribution and value reduction algorithm. Some historical data of a certain power plant are mined by the introduced method, then the attribution and value reduction algorithms are both proved by the correct regulation rules.Finally, the algorithms are further applied in a 300 MW utility boiler in the load of 200 MW. At the first step, confirm the NO_x emission related optimization parameters for mining condition attributes, namely for mining decision table. After preprocessed and dispersed, the data mined by the attribute reduction algorithm and value reduction algorithm. After theoretical analysis, the 11 properties, include flue gas temperature, nozzles tilting angle, secondary air distribution pattern, flow of forced and induced draft fans, shows a strong correlation relationship between these reduced attributes and the emission of NO_x. And the analyzed results show that, in the load of 200 MW, the lower ratio of underlying secondary air, and the higher-level of upper secondary air, and the lower swing angle of burner, low excess air ratio, and the lower exhaust gas temperature can effectively control the NO_x emissions a lower level. Choose the rules’ NO_x emission level next to the lowest can decrease the NO_x emission and maintain the boiler efficiency, which can provide guidance for power plant boiler optimal operation.