The slime is a kind of solid garbage produced in the process of coal washing

which has the characteristics of high viscosity

high moisture content and low calorific value. As a clean and effective combustion technology

circulating fluidized bed(CFB) provides a good method to deal with the slime. The bed temperature of CFB boiler is an important factor influencing the operation security. Mixing slime in CFB boilers decreases the bed temperature and aggravates the bed temperature fluctuation at boundary value. Therefore

the problem of modeling on bed temperature of coal slime co-combustion CFB boiler was studied. On the basis of the running data of actual unit

a dynamic model of the bed temperature was established by least squares support vector machine(LSSVM)dynamic modeling algorithm. The coal-feed quantity

slime quantity

primary air volume

secondary air volume

opening of external heat exchanger and limestone-feed quantity were selected as input variables

and the bed temperature of the CFB boiler was selected as output variable. The simulation results prove that the model has high accuracy

good real-time

and is valuable for bed temperature online prediction.