Soft-Sensor Modeling on NOx Emission of Power Station Boilers Based on Least Squares Support Vector Machines

The online monitoring for NO
_x
emission of coal-fired boilers in power plants is more difficult to achieve. The soft-sensor technology of artificial neural network (ANN) method that was commonly used has not strong generalization ability, but support vector machine modeling-method can solve the problem better. In this paper, a soft-sensor modeling on NO
_x
emission of power station boilers based on least squares support vector machines (LS-SVM) was built. The model can predict NO
_x
emission in different conditions. The comparative analysis of forecast-results between LS-SVM model and ANN model showed that LS-SVM has more strong generalization ability and higher calculation speed.