Levinskyi M. Adaptive control of technological type control objects: algorithms of controllers gain passive self-tuning

The new solution to the topical scientific-technical task of the technological type control object (CO) with varying in wide range gain operation effectiveness increase due to improvement of automatic control system (ACS) self-tuning algorithms was presented in the thesis. Well-known concept of controller self-tuning in closed-loop ACS on the basis of CO passive identification has obtained further development by separating the components of own motion from CO and its model controlled variables total motion, implementing the self-tuning loop, in which the task of CO gain identification is limited to the task of control and the approximate equality of CO and its model controlled variables stochastic characteristics estimates is ensured. The order selection of the Butterworth band pass filters differential equations for separating the CO and its model controlled variables own motion has been substantiated as a compromise between the filtering efficiency and the complexity of their software implementation. New structures and algorithms of self-tuning loops were developed, which compensate shortcomings of known system and increase accuracy, speed, stability and robustness of self-tuning ACS (STACS) with technological type CO which has gain that vary in wide range. STACS self-tuning algorithms for CO startup modes, that allow to complete the self-tuning procedure during the startup period, to prevent unacceptable dynamic control errors and to ensure the unstressed transition of control and self-tuning loops from initializing to steady mode of operation have been developed. STACS control loop algorithms have been improved, which eliminates the steady-state components of control errors, that are proportional to the CO gain rate change. It is shown that the joint application of the developed directions for the improvement of STACS has synergistic effect. Self-tuning algorithms retain their high efficiency for objects with different types of nonlinearities. A technique for determining the parameters of the STACS self-tuning loop has been developed which makes it possible to do calculations using the same initial data as for controllers in the conventional ACS. Software module for Simatic industrial controllers has been created. It implements the algorithms of STACS self-tuning loop and is easily integrated with the functional control units included in the standard libraries of the Step 7 programming environment.