O2 Concentration Measurement of Furnace

Oxygen content of flue is key parameter to judge whether furnace is working on optimal state or not. Long time-lag of O2 Concentration and short life of its detection device have been influenced optimization operation of furnace. A novel Elman Neural Network model is proposed for measuring oxygen content of furnace. The model uses a novel category method to design input parameter of Elman NN, which reduce numbers of input parameter of neural network; therefore it meet the challenge of real-time control. By selecting different time of input value and output value (measured value of oxygen content) to study and train neural network, which time-interval is delay time of oxygen content of flue between the influencing factors of O2 Concentration, soft sensing of oxygen content of flue is changed to chamber´s. Comparing with the data measured by routine device that installed bottom of flue, Trial results show that good dynamic regulation performance of system can be obtained, and fuel efficiency is improved greatly.