Download fulltext HTML

https://doi.org/10.18502/espoch.v3i3.16614

statistics

  • 4 Abstract Views
  • 1 PDF Views
  • 0 HTML Views

authors

  • Hugo Mera

    Hugo Mera

    Investigador independiente, Escuela Superior Politécnica de Chimborazo (ESPOCH), 060155, Riobamba, Ecuador.
  • Félix García

    Félix García

    merahugo1d@gmail.com
  • Edisson Calderón

    Edisson Calderón

    Investigador independiente, Escuela Superior Politécnica de Chimborazo (ESPOCH), 060155, Riobamba, Ecuador.

Published Date

  • Jul 24, 2024

Development of a Method for Diagnosing Faults in Hydraulic Systems Using Artificial Neural Networks with Deep Learning

ESPOCH Congresses: The Ecuadorian Journal of S.T.E.A.M. / Volume 3, Issue 3 / Pages 34–59

Abstract

The application of artificial intelligence is a recent improvement in the industry, allowing preventive maintenance to be applied as a reliability method for detecting failures in hydraulic systems. This is achieved by using artificial neural networks (ANN) as classifiers to make automatic binary and categorical decisions. Since these systems have multiple conditions and sub-conditions that can be complex for normal analysis, the UCI repository database is relied upon to construct an intelligent algorithm of artificial neural networks with deep learning. This has proven to be a highly effective way of predicting failures, with an overall accuracy rate of 97% when applying the intelligent model to the system. As a result, it can be concluded that deep learning is much more efficient than classical machine learning.


Keywords: artificial intelligence, predictive maintenance, artificial neural networks, deep learning.


Resumen


La aplicación de la inteligencia artificial es la nueva mejora en la industria, permitiendo que el mantenimiento preventivo se aplique como método de confiabilidad para la detección de fallos en sistemas hidráulicos aplicando Redes neuronales artificiales (ANN), utilizándoles como clasificadores para obtener una toma de decisiones automáticas de manera binaria y categórica, ya que dichos sistemas poseen varias condiciones y subcondiciones que se vuelven complejas para un análisis normal, apoyándose en la base de datos del repositorio de la UCI, siendo analizados para la construcción de un algoritmo inteligente de redes neuronales artificiales con Deep Learning (aprendizaje profundo), demostrando así un alto desenvolvimiento en la predicción de fallos, obteniéndose un 97% de exactitud (accuracy) de manera general en la aplicación del modelo inteligente al sistema. Se concluye que la aplicación del aprendizaje profundo es mucho más eficiente comparado con el aprendizaje automático clásico.


Palabras Clave: Inteligencia artificial, mantenimiento predictivo, Redes Neuronales Artificiales, Aprendizaje profundo.

References
[1] Prakash J, Kankar PK. Health prediction of hydraulic cooling circuit using deep neural network with ensemble feature ranking technique [Internet]. Measurement. 2020 Feb;151:107225. [cited 2021 Aug 30] Available from: https://www.sciencedirect.com/science/article/abs/pii/S0263224119310905

[2] Guo P, Wu J, Xu X, Cheng Y, Wang Y. Health condition monitoring of hydraulic system based on ensemble support vector machine. 2019 Progn Syst Heal Manag Conf PHM-Qingdao 2019 [Internet]. 2019 Oct 1 [cited 2021 Nov 25]; Available from: https://ieeexplore.ieee.org/abstract/document/8942981 https://doi.org/10.1109/PHMQingdao46334.2019.8942981

[3] Pavlenko I, Trojanowska J, Ivanov V, Liaposhchenko O. Parameter identification of hydro-mechanical. Int J Mechatronics Appl Mech. 2018;(5):19–26.

[4] Zadka M. Installing python. DevOps in Python [Internet]. 2019 [cited 2021 Nov 25];1–6. Available from: https://link.springer.com/chapter/10.1007/978-1-4842-4433- 3_1 https://doi.org/10.1007/978-1-4842-4433-3_1.

[5] MLSys Proceedings. TensorFlow.js: Machine learning for the web and beyond [Internet]. Machine Learning for the Web and Beyond. 2022 [cited 2022 Feb 28]. Available from: https://proceedings.mlsys.org/paper/2019/hash/ 1d7f7abc18fcb43975065399b0d1e48e-Abstract.html

[6] Dua D and G. © UCI Machine Learning Repository [Internet]. Condition monitoring of hydraulic systems Data Set. 2017 [cited 2021 Dec 17]. Available from: https://archive.ics.uci.edu/ml/datasets/Condition+monitoring+of+hydraulic+systems

[7] Rolon-Mérette D, Ross M, Rolon-Mérette T, Church K. Python for research in psychology introduction to Anaconda and Python: Installation and setup. 2020 [cited 2021 Nov 22];16(5). Available from: https://www.spyder-ide.org/

[8] Zhang XD. Machine Learning. A Matrix Algebr approach to Artif Intell [Internet]. 2020 [cited 2021 Nov 29];223–440. Available from: https://link.springer.com/chapter/10.1007/978-981-15-2770-8_6

[9] Hao X, Zhang G, Ma S. Deep Learning. [Internet]. 2016 Nov 30 [cited 2022 Feb 28];10(3):417–439. Available from: https://www.worldscientific.com/doi/abs/10.1142/S1793351X16500045 https://doi.org/10.1142/S1793351X16500045

[10] Dai J, Tang J, Huang S, Wang Y. Signal-based intelligent hydraulic fault diagnosis methods: Review and prospects. Chinese J Mech Eng (English Ed [Internet]. 2019;32(1). Available from: https://doi.org/10.1186/s10033-019-0388-9

[11] Tablada -Germán CJ. Torres A. Redes Neuronales Artificiales. 2021 Aug 10 [cited 2021 Dec 16]; Available from: https://revistas.unc.edu.ar/index.php/REM/article/view/10280

[12] Rivas W, Bertha A, Olivo M. Redes neuronales artificiales aplicadas al reconocimiento de patrones. 2017 [cited 2022 Feb 3]; Available from: www.utmachala.edu.ec

[13] The Theano Development Team. Al-Rfou R, Alain G, Almahairi A, Angermueller C, Bahdanau D, et al. Theano: A Python framework for fast computation of mathematical expressions. 2016 May 9 [cited 2022 Feb 28]; Available from: http://arxiv.org/abs/1605.02688

[14] Manaswi NK. Understanding and working with Keras. Deep Learn with Appl Using Python [Internet]. 2018 [cited 2022 Feb 28];31–43. Available from: https://link.springer.com/chapter/10.1007/978-1-4842-3516-4_2 https://doi.org/10.1007/978-1-4842-3516-4_2

[15] Alhassan AM, Zainon WM. Brain tumor classification in magnetic resonance image using hard swish-based RELU activation function-convolutional neural network. Neural Comput Appl 2021 3315 [Internet]. 2021 Jan 11 [cited 2022 Mar 2];33(15):9075– 9087. Available from: https://link.springer.com/article/10.1007/s00521-020-05671-3 https://doi.org/10.1007/s00521-020-05671-3

[16] Pan SP, Li ZF, Huang YJ, Lin WC. FPGA realization of activation function for neural network. Proc - 2018 7th Int Symp Next-Generation Electron ISNE 2018 [Internet]. 2018 Jun 22 [cited 2022 Mar 2];1– 2. Available from: https://ieeexplore.ieee.org/abstract/document/8394695 https://doi.org/10.1109/ISNE.2018.8394695

[17] Wei Z, Arora A, Patel P, John L. Design space exploration for softmax implementations. Proc Int Conf Appl Syst Archit Process [Internet]. 2020 Jul 1 [cited 2022 Mar 2];2020-July:45–52. Available from: https://ieeexplore.ieee.org/abstract/document/9153236 https://doi.org/10.1109/ASAP49362.2020.00017

[18] Follow WK. Overrtting vs. Underrtting: A complete example. 2018 [cited 2021 Dec 21]; Available from: https://towardsdatascience.com/overfitting-vs-underfittinga- complete-example-d05dd7e19765

[19] Bashir D, Monta nez GD, Sehra S, Segura PS, Lauw J. An Informationtheoretic perspective on overfitting and underfitting. Lect Notes Comput Sci (including Subser Lect Notes Artif Intell Lect Notes Bioinformatics) [Internet]. 2020 Nov 29 [cited 2021 Dec 21];12576 LNAI:347– 58. Available from: https://link.springer.com/chapter/10.1007/978-3-030-64984-5_27 https://doi.org/10.1007/978-3-030-64984-5_27

[20] Babel V, Kumar Singh B, Kumar Jangir S. Journal of Analysis and Computation ( JAC) Evaluation Methods for Machine Learning. 2019 [cited 2021 Dec 15]; Available from: www.ijaconline.com

[21] Xu J, Zhang Y, Miao D. Three-way confusion matrix for classification: A measure driven view. Inf Sci (Ny) [Internet]. 2020 Jan 1 [cited 2021 Dec 16];507:772–94. Available from: https://www.sciencedirect.com/science/article/abs/pii/S0020025519306024 https://doi.org/10.1016/j.ins.2019.06.064.

[22] Zeng G. On the confusion matrix in credit scoring and its analytical properties. https://doi.org/https://doi.org/101080/0361092620191568485. [Internet]. 2019 May 2 [cited 2021 Dec 17];49(9):2080–93. Available from: https://www.tandfonline.com/doi/abs/10.1080/03610926.2019.1568485r-limb prosthesis. RIAI - Revista Iberoamericana de automática e informática Industrial [Internet]. 2008 [cited 2023 Jan 22];5(2):60–68. Available from: http://riai.isa.upv.es

[23] Rosales, Daniela (INSHT) IN de S e H en el T. Evaluación de las condiciones de trabajo: carga postural. Método REBA (Rapid Entire Body Assessment). Instituto Nacional de Seguridad e Higiene en el trabajo [Internet]. 2001 [cited 2023 Jan 22];7. Available from: http://www.insht.es/InshtWeb/Contenidos/Documentacion/ FichasTecnicas/NTP/Ficheros/601a700/ntp_601.pdf

[24] Boné Pina MJ. Método de evaluación ergonómica de tareas repetitivas, basado en simulación dinámica de esfuerzos con modelos humanos. 2016 [cited 2023 Jan 22];268. Available from: https://dialnet.unirioja.es/servlet/tesis?codigo=78749&info= resumen&idioma=SPA

[25] Asensio cuesta S, Bastante Ceca M, Diego Mas J. EVALUACIÓN ERGONÓMICA DE PUESTOS DE TRABAJO - ASENSIO CUESTA, SABINA, BASTANTE CECA, MARÍA JOSÉ, DIEGO MAS, JOSÉ ANTONIO - Google Libros [Internet]. 2012 [cited 2023 Jan 22]. p. 1–350. Available from: https://books.google.com.ec/ books?hl=es&lr=&id=v5kFfWOUh5oC&oi=fnd&pg=PR15&dq=1.+Asensio-Cuesta+ S,+María+José+Bastante+Ceca,+Diego+A.+EVALUACIÓN+ERGONÓMICA+DE+ PUESTOS+DE+TRABAJO.+Editorial+Paraninfo%3B+2012&ots=wJVRmKqsFL&sig= Zzl4RNRdOO5Zwb11n4DXijNbeUE#v

[26] Guerrero Silva CF. Universidad de Guayaquil. Facultad de Ingeniería Industrial. Carrera de Ingeniería Industrial. Universidad de Guayaquil. Facultad de Ingeniería Industrial. Carrera de Ingeniería Industrial.; 2019 [cited 2023 Jan 22]. p. 72 Repositorio Universidad de Guayaquil: Evaluación de riesgo ergonómico aplicando el método REBA a los trabajadores administrativos de la EmpresaPública Municipal Registro de la Propiedad de Guayaquil. Available from: http://repositorio.ug.edu.ec/handle/redug/42144