Applications of neural network in the oxygen content prediction of boilers´ smoke

On the background of the inefficient burning systems of boilers, neural network modeling techniques were studied intensively and presented to predict the oxygen content of boilers´ smoke online. This pointed out a new way for improving the control of the burning systems. The problems arising from implementing a neural network model in real industrial processes are also considered and solved, including data collection, data pretreatment, soft model correction, and so on.