Temperature of flue gas exhausted from utility boilers is an improtant index affecting the heat loss of boilers

and it is comprehensively affected by many factors

such as boiler load

coal sort

and controlled parameters in the combustion system etc.

the relationship among them is difficult to be described by using a linear mathematical formula.By using the historical operation data of boiler

as well as the data of furnace-entering coal sort

the temperature model of flue gas exhausted from boilers has been established by adopting three nonlinear tools

i.e.the support vector machine(SVM)

least square support vector machine(LSSVM)

and BP neural network(BPNN)

respectively

at the same time

the model's parameters of the first two algorithms have been optimized and selected by using the cross-validation method and combining with the genetic algorithm.Comparatoin and analysis of prodicted results from three prediction models show that the generalization ability of SVM method is comparatively high

and has rapid converging speed.In view of the influence of coal quality upon the combustion condition in boilers

under condition of considering the facter of coal quality

the predictionmodel for temperature of flue gas exhausted from boiler has been established and predicted by usingSVM method respectively

results show that the establishment and prediction of model after considering coal quality boast even higher accuracy.