Font Size: a A A

Study For The Control Of Gauge And Flatness In Cold Rolling And Intelligent Optimization

Posted on:2011-06-17Degree:MasterType:Thesis
Country:ChinaCandidate:L H ZhangFull Text:PDF
GTID:2121360302994804Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
It is important index to gauge precision and good flatness in cold rolling strip. Gauge is controlled by hydraulic pressure. Flatness is controlled by curved roller. The research of gauge control is mature, and there are many methods, for example, feedback control, feed-forward control, stakeout control, tension control, flow control and so on. The research of flatness is slow because flatness measure is difficult. The technology of flatness control is not advanced. The integrated control of gauge and flatness was discussed by many scholars, but it is used less in the actual field of cold rolling. The integrated system of gauge and flatness is coupling, nonlinear, and delay. The field of cold rolling mill is complex, and process of rolling is influenced by different factors, so the difficulty of integrated system is increased.In the background of practical engineering project, this paper firstly analyses the basic theory of gauge and flatness, and shows control systems of cold rolling. The models of the cold rolling hydraulic pressure depress the automatic position control system and the hydraulic automatic roll-bending force servo control system are established and equivalent simplified. The output of step respond is ensured the same after simplified.Secondly, integrated system model of gauge and flatness is established. The model is decoupling by diagonal control and invariance principle. Simulation results of coupling and decoupling are showed. The complex system is also decoupling by RBF neural network and PID control combined. The simulating results prove that the effect of decoupling is well and systems can restrain disturbance.Finally, PID parameters of decoupling integrated system are optimized by MMAS. MMAS has characters of distributed counting and positive feedback. The global optimization is found quickly and search process is not easily stagnant because of local optimization. Results show that step respond outputs optimized by MMAS are quickly and less overshoot.
Keywords/Search Tags:Cold rolling mill, Gauge and flatness control, Decoupling, RBF neural network, MMAS, PID control
PDF Full Text Request
Related items