Steady-state condition extraction is the basis of combustion optimization for utility boiler with historical operational data.Thus

a slippage distinguishing algorithm combining with improved confidence interval was proposed

to find out the steady-state cases among the historical operational data of utility boiler.Moreover

this algorithm is applied to detect the steady-state cases of a subcritical 660 MW unit boiler.The results indicate that

with an increase in the unit load

the NOxemission from this boiler decreases first and then increases.In the same load level

methods aiming at strengthening the effects of staged combustion

such as amplifying the SOFA damper opening degree and regulating the coal mill combination mode

can decrease the NOxemission.Co-firing coal with high volatile and low ash content is helpful to NOxemission reduction.