The 35t/h boiler used in the Thermal Power Plant of Tangshan Iron and Steel Co. Ltd. is a chain furnace. Its control system of the conmbustion process is a 2 input-2 output system whose inputs are the quantity of the fuel and desuperheater spray, outputs are the steam pressure and the steam temperature. Due to the nonlinearity, large lag, and coupling impact between the input and output parameters, fuzzy control was introduced. The fuzzy control rule has the adjustable factors and is described by the analytic relationships. The steam pressure is a 2-level control, including the original and final control. In the final control, the effect of the fuel quantity change on the steam temperature is taken into consideration. Thus the correction of the steam temperature for the fuel quantity is introduced. The control rule can be adjusted on-line owing to the regulable factors. For the reason, the control is closer to the actual production, which makes the production more smooth and reliable. Lastly, the modular structural design by means of PLC makes the realization of the fuzzy control algorithm more convenient.
Automatic generation of fuzzy control rule by machine learning methods