Sintering finish point intelligent control

Sintering process is very important for blaster furnace production. This plant is a heavy nonlinear unknown object, the traditional control is hard to achieve good results. Traditional adaptive pole placement method is effective in linear control system theory with an accurate linear model essentially. Its capability to overcome nonlinear disturbance is limited. The neural network identifier can set up an accurate model for an unknown nonlinear object. A neural network model with a special structure, which is divided into linear and nonlinear parts, is applied into identify an unknown nonlinear system object in this paper. The model identification speed and accuracy are improved. Then, considering the nonlinear part of object as measured disturbance, can be compensated by a feed forward passage, an adaptive pole placement algorithm was investigated in controlling the linear part of object in time. The practical control result on sintering finish point verified the new controller effect