Methodology for Selecting the Discretization Step of a Model in Information and Measurement Technologies

Valerii Mazurenko

ORCID: https://orcid.org/0000-0001-8340-012X

Oles Honchar Dnipro National University

Purpose. The main aim of the research is to produce recommendations in the form of a methodology that permits designers to establish a discretization step for a simulation model, when that model is being prepared to be included into information and measurement technology, as well as sampling rate of information and measurement technology, when that technology is implemented at real-time system of process control. The methodology should be based on the characteristics of the model itself and establish simple, clear, and easy-to-use rules. It should be possible to find discretization steps without physical experiments with real objects. Design / Method / Approach. The research uses theory of signals and systems and control theory. The methodology is based on the spectral characteristics of signals that circulate inside observed/controlled objects and dynamic characteristics of those objects. Findings. The paper presents formulas for calculating discretization step value that is based on value of relative sampling error of signal representation, as well as on base of values of time constants given in the object model. The recommendation on how to adjust it in relation to the real-time system’s main cycle value is represented as well. Theoretical Implications. Results achieved under presented research develop applied methods of control theory. Practical Implications. The proposed method has practical results due to giving useful instruments for real-time systems designers that simplifies processes to define the value of crucial system parameters. Originality / Value. The observed problem is quite actual because there are no sources that present simple and clear methodology to solve it in practice. Research Limitations / Future Research. The research is limited by the category of measurement and control systems that use object model functioning in real-time mode to estimate parameters that are not directly observed.



RELATED PAPERS