Abstract

The authors develop an adaptive approach to synthesis of intelligent digital models, which have acquitted particular significance during the research of controlled mechanical constructions. Capacity for generation of active reactions to external functional and energy impacts is a distinctive feature, when the object automatically finds rational or optimal states by means of respective adaptation to the aforesaid impacts. The paper presents principles of intelligent model construction, which simulates the processes of controlled transitions between various states of "the controlled construction" in presence of NON-factors. The principle of cyclic adaptive control has been taken as the basis of generation of the "controlled construction" intelligent model, where in every cycle the managerial decision generation technology is implemented, which provides maintenance of the current construction state or its transition to a more preferable state. The identification process for identification of states of the studied construction has been analyzed, and it assumes comparison of the current values of its parameters corresponding to microstates of the informational image along with classification parameters of the preliminary designated multiple taxons of microstates. A taxonomic scheme was offered for this purpose based on the use of five main taxons having the definite system meaning. It has been shown that application of the squared distance metrics to median abscissa of the selected intervals is technologically sufficient as a proximity measure of the parameter value to the taxon of microstates. The management scenario has been considered in the intelligent mechatronic system, which displays transitions between the optimal states of "the controlled construction" under the influence of external factors. An example, of the intelligent model of mechanical construction is shown in the form of a pulse-type regulator for control of the executive mechanism of system reactions. Construction of the pulse-type regulator is based on the use of the virtual signal formation scheme of mismatch between the current value of the input signal and its value defined by the intelligent system setting mechanism. This mismatch is subject to discrete convolution with the tabulated and abridged original of the required pulse characteristic. It is also shown that sufficiently economical recurrent interrelations for calculations according to mismatch values are formed during construction of the model on the basis of proportional integral and differential regulators.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call