Research On Soft-Sensing Of Oxygen Content In Flue Gas Of Power Plant

The main way to improve the economy of utility’s operation is improving the efficiency of the boilers. Traditional mode is coarse control the air amount based on the ratio of air and coal, making sure the input air is enough for the combustion. While the working condition is steady, the engineers adjust the ratio of air and coal on the basis of the oxygen content in flue gas, because normally the oxygen content is low when the load is high and high when the load is low. Combustion condition can be judged according to the oxygen content in flue gas. This is the significance of measuring the oxygen content timely and accurately. However, the hardware measurement cannot meet the need of control because of non-real-time and poor reliability. The soft-sensing technique provides a new method of measuring, and it’s useful for combustion optimization control of boiler system.Power plant operation optimization is to both safety and economic indicator, but also to evaluate operation condition of the unit and decision-making, eventually to guide the operation adjustment. Plant there are a large number of data real-time historical database of information, these are very important field of operation information, if the in-depth analysis of these data can find the law of the unit operation and improve the method of system. Based on oxygen for the soft measurement on the basis of mechanism analysis, the main factors affecting the flue gas oxygen content are discussed in this paper. Using correlation of plant-operation data exist, ensure reliable input variable data integrity. Extracted from a60mw coal-fired power station unit stable operation history data, establish neural network soft measurement model of flue gas oxygen content.