Redes neuronales regularizadas aplicadas a la estimación de propiedades mecánicas de perfiles de acero//Regularized neural networks applied to the estimation of the mechanical properties of steel profiles

Authors

  • Doniel Jiménez-Sánchez Universidad Tecnológica de La Habana José A. Echeverría, Centro de Estudios de Matemática. La Habana.

Abstract

El uso de las redes neuronales artificiales de base radial para la estimación de las propiedades mecánicas de las hornadas a partir de su composición química y superficie transversal del perfil terminado,facilita el proceso de generación de opciones de destino de hornadas de perfiles ligeros de acero destinados a la construcción. En el trabajo realizadoesta estimación se hizopor redes de base radial entrenadas con el método iterativo de regularización de Landweber, a partir de los datos acumulados de ensayos mecánicos en los talleresadoptados como Estudio de Caso, los que producen fundamentalmente barras corrugadas de acero para su uso en la construcción. La utilización de redes regularizadas permite disminuir los errores de generalización en la estimación de las propiedades mecánicas, mejorando así la calidad de las opciones de destino de las hornadas producidas. 

Palabras claves: redes neuronales artificiales de base radial, estrategias de regularización, dirección operativa de la producción.
______________________________________________________________________

Abstract

The use of the radial baseartificial neural networks for the estimation of the mechanical properties of the heats starting from its chemical composition and traverse surface of the finished profile facilitates the options generation process forthe heats destination while producingsteel slight profiles dedicated to the constructions. In the present work this estimation is made by radial base artificial neural networks, trained by the Landweber regularization iterative method, starting from the available mechanical properties data obtained from the quality control of the workshops adopted as Study Case. The use of regularized networks allows diminishing the generalized errors in the mechanical properties estimation, improving this way the quality of the options of destination of the heats produced.

Key words: basis neural networks, regularization strategies,operative production management

Downloads

Published

2017-09-02

How to Cite

1.
Jiménez-Sánchez D. Redes neuronales regularizadas aplicadas a la estimación de propiedades mecánicas de perfiles de acero//Regularized neural networks applied to the estimation of the mechanical properties of steel profiles. Ing. Mec. [Internet]. 2017 Sep. 2 [cited 2025 Sep. 13];20(3):115-21. Available from: https://ingenieriamecanica.cujae.edu.cu/index.php/revistaim/article/view/566

Issue

Vol. 20 No. 3 (2017): septiembre-diciembre

Section

Original article

License

Authors who publish with this journal agree to the following terms:

a. Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.

b. 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.

c. 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 (See The Effect of Open Access).