THE PROBLEM OF IDENTIFICATION OF LINEAR STATIONARY OBJECTS WITH DISTRIBUTED PARAMETERS BY THEIR EXPERIMENTAL TRANSIENT CHARACTERISTICS

A wide class of control system elements can be described with reasonable accuracy by the concept of a linear stationary dynamic object. Several mathematical descriptions of such an object are known. The traditional mathematical model is a high-order ordinary linear differential equation. In the Laplace image space, this corresponds to a fractional-rational transfer function. The latter can be decomposed into elementary fractions. Then, using the convolution theorem and tables of elementary Laplace transform functions, one can access the originals. It is crucial to ensure precise alignment of the parameters of the mathematical model of the object used in the corrector with the actual parameters of the object. If this condition is violated (for example, due to sensor aging during operation in an aggressive environment), additional dynamic error arises, which can significantly reduce the effectiveness of the correction. It is important not only to ensure the adequacy of the mathematical model relative to the actual object but also to promptly monitor changes in the object's parameters. To achieve this, effective algorithms for parametric identification of the object model are required. The heuristic algorithm for identifying a linear stationary aperiodic object with distributed parameters based on an experimental transient response is proposed. The time constant and degree of inertia are determined as parameters of the object model in the form of an irrational transfer function.

Перевод пока недоступен