On the basis of test data of coal-fired boiler combustion adjustment in a power plant

a hybrid model of the boiler combustion system was established by using the BP neural network

targeting at predicting the boiler efficiency and NOx emission.Moreover

by the genetic algorithms

the weights and thresholds of the BP neural network model were optimized to obtain the improved GA-BP neural network model.Finally

comparison of the prediction error between above two models was conducted.The results showed that

compared with the BP neural network model

by the GA-BP neutral network model

the average prediction error declines from 6% to 3.4% for boiler efficiency and from 0.1% to 0.04% for NOx emission concentration.The generalization ability of the GA-BP neural network model was greatly enhanced

and this model was suitable to build the boiler combustion optimization system.