Search

The control of surface roughness is of great significance for measuring the quality characteristics of cold-rolled steel strip. At present, the conventional method is used to monitor the surface roughness of steel strip by random sample inspection, off-line analysis and test, and on-line adjustment of production process, but the method is of some hysteretic nature and discontinuity. To this end, the paper presents an on-line intelligent control method for surface roughness of the cold-rolled steel strip, adjusting parameters of the skin-pass process in real time and reducing the fluctuation of surface roughness of strip steel. To establish the fuzzy neural network control model of the corresponding relationship between the parameters of the skin -pass process and the surface roughness of steel strip, the deviation between the online detection value and the target value of the surface roughness is taken as the input of the model, so is the differential value of deviation. In addition, the skin -pass rolling force is used as the output. Learning and training fuzzy neural network to get a fuzzy neural network online prediction control model that is applied to continuous treatment unit for cold rolling strip steel. By measuring the surface roughness of steel strip, the system adjusts the skin-pass rolling force in real time. At the same time, to keep the mechanical properties of the steel strip and the strip shape, the system optimizes skin -pass rolling tension and bending force accordingly to realize on -line intelligent control for the surface roughness of steel strip. The proposed system can improve the control accuracy of surface roughness, which can even better satisfy the requirements of downstream users.