SUMMARY
Technological processes as control objects are characterized by substantial delay in controlled variables response to control action andconsiderable quantity of varying factors, which have influence on the process, but, practically, inaccessible for measurement. These specifics,in practice, often lead to substantial decrease in operation quality of the control system with typical algorithms and significantly make it harderto design and develop self-tuning control systems for these types of control objects. Factors, depending on the sequences of the impacts they have on the control object, are subdivided into non-controllable coordinate and parametric disturbances. In this article the case is considered, when spectral composition of parametric disturbances is substantially lower in frequency range in comparison to coordinate disturbances spectral composition. Factors, which cause high-frequency changes in controlled variables, which cannot be compensated by the control actions, are considered as noises. For control objects of technological type, in which transition coefficient changes under the influence of parametric disturbances, the structure of self-tuning automatic control system is proposed. Unlike many known, principle of its operation includes passive identification of the transition coefficient changes in the closed-loop circuit. From the overall movement of the closed-loop system, which arises under the influence of the coordinate disturbances, the part of the own motion of the system is being separated utilizing band pass filters. By detecting the changes in spectrum of this component, self-tuning loop determines current value of control object transition coefficient and changes the value of controller transition coefficient in the stabilization circuit for maintaining the stable operation of the system. Computer experiments for evaluation of how coordinate disturbances and noises spectral composition change influence the quality of self-tuning were carried out, the possibility of carrying out the optimal parametric synthesis of the system is shown, recommendations for approximate evaluation of the self-tuning loop parameters are proposed.