REFINEMENT AND ACCURACY CONTROL OF THE SOLUTION METHOD FOR THE DURABILITY PROBLEM OF A CORRODING STRUCTURE USING NEURAL NETWORK
DOI:
https://doi.org/10.15588/1607-3274-2024-1-9Keywords:
artificial neural networks, accuracy control, distribution, mathematical expectation, approximation, numerical methods, durability corroding structureAbstract
Context. The prediction of the time until failure of corroding hinge-rod structures is a crucial component in risk management across various industrial sectors. An accurate solution to the durability problem of corroding structures allows for the prevention of undesired consequences that may arise in the event of an emergency situation. Alongside this, the question of the effectiveness of existing methods for solving this problem and ways to enhance them arises.
Objective. The objective is to refine the method of solving the durability problem of a corroding structure using an artificial neural network and establish accuracy control.
Method. To refine the original method, alternative sets of input data for the artificial neural network which increase information about the change in axial forces over time are considered. For each set of input data a set of models is trained. Based on target metric values distribution among the obtained sets, a set is selected where the minimum value of the mathematical expectation of the target metric is achieved. For the set of models corresponding to the identified best set, accuracy control of the method is determined by establishing the relationship between the mathematical expectation of the target metric and the parameters of the numerical solution.
Results. The conditions under which a lower value of the mathematical expectation of the target metric is obtained compared to the original method are determined. The results of numerical experiments, depending on the considered case, show, in average, an improvement on 43.54% and 9.67% in the refined method compared to the original. Additionally, the proposed refinement reduces the computational costs required to find a solution by omitting certain steps of the original method. An accuracy control rule of the method is established, which allows to obtain on average a given error value without performing extra computations.
Conclusions. The obtained results indicate the feasibility of applying the proposed refinement. A higher accuracy in predicting the time until failure of corroding hinge-rod structures allows to reduce the risks of an emergency situation. Additionally, accuracy control enables finding a balance between computational costs and the accuracy of solving the problem. KEYWORDS
References
Khoma М. С. State and Prospects of Research Development in the Field of Corrosion and Corrosion Protection of Construction Materials in Ukraine: According to the Materials of Report at the Meeting of the Presidium of NAS of Ukraine, October 27, 2021, Visn. Nac. Akad. Nauk Ukr., 2021, pp. 99–106. DOI: https://doi.org/10.15407/visn2021.12.099
Zelentsov D. G., Korotka L. I., Denysiuk O. R. The Method of Correction Functions in Problems of Optimization of Corroding Structures, Advances in Computer Science for Engineering and Education III. ICCSEEA 2020. Advances in Intelligent Systems and Computing, 2020, Vol. 1247, pp. 132– 142. DOI: https://doi.org/10.1007/978-3-030-55506-1_12
Zelentsov D. G., Korotkaya L. I. Use of neural networks at the decision of problems durability of designs subject to corrosion, Transactions of Kremenchuk Mykhailo Ostrohradskyi National University, 2011, № 3 (68), Part 1, pp. 24–27.
Denysiuk O. R. Determination of rational parameters for the numerical solution of systems of differential equations, Visnyk of Kherson National Technical University, 2016, № 3(58), pp. 208–212.
Zelentsov D. G., Korotkaya L. I. Technologies of computational intelligence in problems of modeling dynamic systems, Monograph. Balans-Klub, 2018, P. 179. DOI: 10.32434/mono-1-ZDG-KLI
Rudenko V. M. Matematychna statystyka. Navch. Posib. Kyiv, Tsentr uchbovoi literatury, 2012, 304 p.
Hecht-Nielsen R. Kolmogorov’s mapping neural network existence theorem, IEEE First Annual Int. Conf. on Neural Networks, 1987, Vol. 3. pp. 11–13.
Goyal M., Goyal R., Venkatappa Reddy P., Lall B. In: Pedrycz W., Chen S.-M. (eds) Activation Functions, Deep Learning: Algorithms and Applications. Studies in Computational Intelligence, 2020, Vol. 865, pp. 1–28. DOI: https://doi.org/10.1007/978-3-030-31760-7_1
Riedmiller M., Braun H. A direct adaptive method for faster backpropagation learning: The RPROP algorithm, Proc. IEEE Int. Conf. Neural Netw., 1993, pp. 586–591. DOI: 10.1109/ICNN.1993.298623
Cook J. D. Inverse gamma distribution. Online: http://www. johndcook. com/inverse gamma. pdf, Technical. Report, 2008.
Ketkar N., Moolayil J. Introduction to PyTorch. In: Deep Learning with Python. Apress, Berkeley, CA, 2021, pp. 27– 91. DOI: https://doi.org/10.1007/978-1-4842-5364-9_2
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2024 O. D. Brychkovskyi
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Creative Commons Licensing Notifications in the Copyright Notices
The journal allows the authors to hold the copyright without restrictions and to retain publishing rights without restrictions.
The journal allows readers to read, download, copy, distribute, print, search, or link to the full texts of its articles.
The journal allows to reuse and remixing of its content, in accordance with a Creative Commons license СС BY -SA.
Authors who publish with this journal agree to the following terms:
-
Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License CC BY-SA that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
-
Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
-
Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work.
