Neural Internal Model Control for Lead-zinc Sintering Feeding Process
|School||Central South University|
|Course||Control Science and Engineering|
|Keywords||sintering feeding single neuron internal model control nonlinear and time delay systems|
Sintering is an important part in the lead and zinc smelting process, it blends the raw material, solvent and fuel with different components, according to the certain ratio to carry on the blending and obtain the sintering blending minerals. The basic requirement of the sintering is feeding precisely, That is feeding continuously and stably in accordance with the calculated ratio, thus can limit the fluctuation of the actual feeding amount within the allowable range, otherwise it would cause the changes of sinter strength and reducibility.Based on the detailed analysis of the site process and the model of the sintering feeding process, the transfer function of feed object is derived. Due to the inherent characteristics of the sintering feeding process and the influence by many uncertain factors in the site, the control process has the characters of high nonlinearity, large lag and varying of the model parameters.Based on these characters, the neural internal model controller is proposed. It has the characteristics of dual stability, ideal controller and zero steady state error, so it can be a good resistance to interference and solve the problem of model mismatch. At the same time, it can eliminate the delay and the adverse effects of the sintering feeding control system. Using the characteristics of single neuron, the filter time constant of the internal model controller is adjusted on the real time. This approach enhances the ability of the internal model control to adapt to changes of the controlled object.By the simulation experiment, the control strategy is effective, and this control method can achieve not only good dynamic and static performance, but also strong robustness and self-study capability simultaneously. The sintering feeding automatic control system is developed. The real-time monitor of feeding process parameters and the technological process are realized. And the optimal control of the feeding process can ensure the smooth operation of the sintering process. Thus the sinter quality and quantity are improved, while the labor intensity of the operator is reduced.