Identification of an object model in the presence of unknown disturbances with a wide frequency range based on the transition to signal increments and data sampling

Ryabchikov M.Y., Ryabchikova E.S.
 pdf (733K)

The work is devoted to the problem of creating a model with stationary parameters using historical data under conditions of unknown disturbances. The case is considered when a representative sample of object states can be formed using historical data accumulated only over a significant period of time. It is assumed that unknown disturbances can act in a wide frequency range and may have low-frequency and trend components. In such a situation, including data from different time periods in the sample can lead to inconsistencies and greatly reduce the accuracy of the model. The paper provides an overview of approaches and methods for data harmonization. In this case, the main attention is paid to data sampling. An assessment is made of the applicability of various data sampling options as a tool for reducing the level of uncertainty. We propose a method for identifying a self-leveling object model using data accumulated over a significant period of time under conditions of unknown disturbances with a wide frequency range. The method is focused on creating a model with stationary parameters that does not require periodic reconfiguration to new conditions. The method is based on the combined use of sampling and presentation of data from individual periods of time in the form of increments relative to the initial point in time for the period. This makes it possible to reduce the number of parameters that characterize unknown disturbances with a minimum of assumptions that limit the application of the method. As a result, the dimensionality of the search problem is reduced and the computational costs associated with setting up the model are minimized. It is possible to configure both linear and, in some cases, nonlinear models. The method was used to develop a model of closed cooling of steel on a unit for continuous hot-dip galvanizing of steel strip. The model can be used for predictive control of thermal processes and for selecting strip speed. It is shown that the method makes it possible to develop a model of thermal processes from a closed cooling section under conditions of unknown disturbances, including low-frequency components.

Keywords: identification, big data, global model, increments, unknown impacts, data sampling
Citation in English: Ryabchikov M.Y., Ryabchikova E.S. Identification of an object model in the presence of unknown disturbances with a wide frequency range based on the transition to signal increments and data sampling // Computer Research and Modeling, 2024, vol. 16, no. 2, pp. 315-337
Citation in English: Ryabchikov M.Y., Ryabchikova E.S. Identification of an object model in the presence of unknown disturbances with a wide frequency range based on the transition to signal increments and data sampling // Computer Research and Modeling, 2024, vol. 16, no. 2, pp. 315-337
DOI: 10.20537/2076-7633-2024-16-2-315-337
URL: http://crm-en.ics.org.ru/journal/article/3458/
Creative Commons License This work is licensed under a Creative Commons Attribution-NoDerivs 3.0 Unported License.

Indexed in Scopus

Full-text version of the journal is also available on the web site of the scientific electronic library eLIBRARY.RU

The journal is included in the Russian Science Citation Index

The journal is included in the RSCI

International Interdisciplinary Conference "Mathematics. Computing. Education"

Copyright © 2009–2024 Institute of Computer Science