Automated control of water-air cooling process for sheet products quality improving

Abstract

An analysis of the main existing methods for modelling objects of the automated control is executed in terms of their efficiency: adequacies of control parameters forecasting as well as the necessary amount of calculating operations per a unit time or duration of the calculations. The necessity to use, in general case, the systems of many complicated multicomponent equations of regressive character to achieve adequate determination of evolution of the most real objects has been shown. In order to increase the automated control efficiency a new physically founded method for multilevel modelling the real objects with multidimensional objective functions has been offered. The method consists of: determination of «new» objective function which has close cross-correlation with every initial parameter; development of systems of regressive models relating to a high adequacy a «new» objective function to each control parameter; development of highly adequate regressive model relating a «new» objective function to each entry parameter. Verification of efficiency of the offered method was executed by the modelling the process of the water-air cooling (WAC) for sheets under the conditions of successive WAC and with the use of air-water mixture. A system of the universal physically founded regressive models has been developed for the rolled steels mechanical properties indexes which under the conditions of successive WAC need with the use of the known methods at least 18 iterations to calculate every separate value of the control parameters while according to the offered approach, the proper amount is just 9 iterations and higher forecasting adequacy has been reached