Experimental Parametrization of the Dual-Mass Electromechanical System of a Rolling Mill

Abstract

Transients in electromechanical systems need to be modeled adequately if the existing electric drive control algorithms are to be improved, or new ones are to be developed. This fully applies to the electric drives of roll-ing mill stands. When studying such drives, one often needs to find a tradeoff between the complexity of Park-Gorev equation-based mathematical description of processes in variable frequency drives, and the capabilities of simpler models that adequately describe the processes relevant for each specific case. The latter include the known dual-mass model of the electromechanical system of a rolling mill stand featuring a flexible shaft and gaps in transmissions. Research presented herein covers the parameters of this model, which need to be ob-tained experimentally. The most accurate way to find the model parameters is to derive them from the oscillo-grams sampled on the object when running in typical transient. This particular paper addresses the electrome-chanical system of the horizontal stand in Plate Mill 5000 deployed at Magnitogorsk Iron and Steel Works (MMK PJSC). The paper substantiates the methodology behind parametrizing the system and shows how to cal-culate the moments of inertia of rotating masses from oscillograms of unloaded drive acceleration. To showcase the method, the paper presents finding the stiffness factor of an elastic coupling from the transients recorded when halting the electric drive. Focus is made on finding the time constants of the internal torque control cir-cuit as approximated by first-order or second-order filters. By comparing the simulation output against the os-cillograms sampled on the mill itself, the model was proven to adequately compute the parameters of the object. The paper concludes with recommendations on how the developed methods could be used to find the parame-ters of a dual-mass system when modeling the electric drives in rolling mills. The conclusions outline possible further research.