The Design Of Brief Process Control Level In900mm Temper Mill

The temper mills exit in most plate and strip production.It is the last but crucialprocess befor the plate and strip being further processed. The temper mill can effectivelyimprove the shape, flatness and drawability of the plate and strip. In order to meet therequirements of users about strip shape, mechanical properties, surface quality and so on,steel mills mostly equipped the temper mill with computer system of production processcontrol level (level2).The brief process control level in900mm temper mill was developed using VB6.0forthe reason of that the rolling procedure of temper mill is much simpler than cold rolling.The system determines the rolling parameter settings (such as: elongation, decoil tension,coiling tension, rolling force, operating speed) by the means of enquiring the processesforms. In addition to the function of querying process, the system has also developed fourauxiliary functions: Login user management, coils management, rolls management,reports and forms generation. The auxiliary functions modules will help to manage andresearch the temper mill better.In order to better achieve the function of rolls management, the system not only hasconfigured some routine functions, like adding, editing and deleting rolls information,querying rolls information which can replace the rolls in use.The system but also hasadded a attenuation model of working rolls’ surface roughness to predict the working roll’ssurface roughness accurately by analyzing roll wear mechanisms. Firstly, the factorswhich influence the working roll’s surface roughness has been analyzed by the means ofGrey Relational Analysis. A system for evaluating the working roll’s surface roughness isdetermined. Then optimized OS-LSSVR model is used for on-line prediction of thesurface roughness. The new key nodes are added recursively by using prediction errorcriterion, and the redundant key nodes are deleted following FLOO. Moreover, thegradient descent method is adopted to optimize the two hyper-parameters online. Theresults of simulation show that the average absolute error of the model is0.0149, is muchsmaller than other models. In addition, the model has on-line adaptive ability, and is able to evolve over time.The brief process control level has acquired favorable effect after it has been appliedin a steel mill for fourteen months.