Application Research Of Data Mining On The Strip Cold Rolling

With the advance of China’s modernization，modern industrial production also putforward higher requirements on the thickness accuracy of strip products, at the sametime, the contradiction between the demand for high value-added cold-rolledproduction in modern industrial production and domestic development level ofprecision strip production status, so research to improve the cold-rolled strip finishedthickness accuracy is important. The control of cold ironing is a multivariable process.Large amounts of data was generated during the rolling process, yet most of whichwere only stored in the database instead of being applied properly. Therefore thispaper has provided an method of data mining in cold rolling. The internal law and thechanging process could be found and predicated by analyzing these huge amounts ofdata, thus to compensate for the outcome to increase the accuracy of slab thickness.This paper has studied the data mining algorithm used in research and the featuresof linear regression and wavelet analysis were discussed.The algorithm was used in determining the roller rigidity. Apart from regular dataprocessing, the efficient data mining software Clementine12.0was used to run theanalysis which derived the mathematical model of roller rigidity and whose accuracywas confirmed.The algorithm was used in determining roll eccentricity model. The collectedsignal was firstly extracted using wavelet analysis and then secondary extracted usingFuriour transformation to get the roll eccentricity signal with higher accuracy. Theexperiment on mill roller300has achieved certain effect which laid the foundation forfurther research.