An improved modeling method for soft-sensing of oxygen content in flue gas and the simulation

Aiming at the importance of oxygen content in flue gas at tail part of the boiler to combustion efficiency, a soft-sensing model of oxygen content was set up through mechanism analysis. Following modifications were carried out for the model: adopting history-data-fitting algorithm to modify the total air flow in furnace; using amplitude limiting filter for filtration of boiler drum pressure in heat signals to improve the dynamic response of soft-sensing process for oxygen content; determining the optimum excess air coefficient under standard state required for per unit heat by data digging method. Simulation was performed according to the history data of a certain power plant. According to indexes of variance, correlation coefficient and maximum deviation, it is proved that the improved soft-sensing model of oxygen content is effective.