The pulverizer mathematicalmodel in (1) is described by several differential and algebraicequations as follows:

M_c mass of coal in pulverizer
M_pf mass of pulverized coal in pulverizer
T_out outlet temperature of pulverizer
△P_mill pulverizer differential pressure
W_pf mass flow rate of pulverized coal
P pulverizer power consumed
△P_pa primary air differential pressure
W_c mass flow rate of coal into pulverizer
T_in inlet temperature of pulverizer
W_air primary air flow into pulverizer
K _i the constant coefficients of the pulverizer model,i=1,2,3,…,12.

the whole mathematic model ofa pulverizer can be obtained after identifying the coefficients inequation set (1)by the improved GA. However, it is not convenient todesign the control system of the pulverizer .so the above modelsystem of the pulverizer. So the above model needs to be convertedinto a T-S fuzzy model baese on analyzing the operation mechanism. Tominimize the design effort, we try to use as few rules as possiblewith guarantee of the accuracy of the T-S fuzzy model. Let theprimary air differential pressure △P_pa and the primaryair-flow into pulverizer W_air be the premise variables ofT-S fuzzy model. Wair and the product of T_in and W_airT_in W_air are the inout variables. W_pf ,△P_mill, Tout, P-K is the output variables. Then thenonlinear model can be converted into an equivalent T-S fuzzy modelof the pulverizer as follows:

IF △P_pa is P_iand W_air is W_i

Blending linear models (1)-(4)through the above nonlinear fuzzy membershio functions μ_i^j =(i=1,2,j=1,2,3,4) , the T-S fuzzy model of thepulverizer can be obtained as follows:

it is composed of four linearmodels that are blended through nonlinear membershio functions.Controllers with good performance to every linear system and thewhole nonlinear system can be designed based on the T-S fuzzy modeleasily. So the T-S fuzzy model is suitable for designing the controlsystem of the pulverizer.