Classification of Mechanical Failures in Provoked Ignition Engine by Means of ANN and SVM

Rafael Wilmer Contreras Urgilés, José Maldonado Ortega, Esteban Rocano, Jorge Chiluisa

Producción científica: Capítulo del libro/informe/acta de congresoContribución de conferenciarevisión exhaustiva

Resumen

This paper presents the methodology applied to determine the mechanical failures in an internal combustion engine caused by the application of artificial intelligence in the classification of mechanical failures associated with the cancellation of cylinder work, that is to say this methodology is applied on the data obtained from the signal of the KS sensor (Knock Sensor) and the CMP sensor (Camshaft Position Sensor) during engine operation. To evaluate the data obtained, the acquisition of samples applied to different operating conditions is carried out, after which an attribute matrix is created that allows a selection and reduction of variables with the application of methods based on the Random Forest architecture. Subsequently, an ANN (artificial neural network) and an SVM (support vector machine) was created and trained, from which a classification error value of 0.1267% and 0.0067%, respectively, was obtained.

Idioma originalInglés
Título de la publicación alojadaIntelligent Technologies
Subtítulo de la publicación alojadaDesign and Applications for Society - Proceedings of CITIS 2022
EditoresVladimir Robles-Bykbaev, Josefa Mula, Gilberto Reynoso-Meza
EditorialSpringer Science and Business Media Deutschland GmbH
Páginas161-172
Número de páginas12
ISBN (versión impresa)9783031243264
DOI
EstadoPublicada - 2023
Evento8th International Conference on Science, Technology and Innovation for Society, CITIS 2022 - Guayaquil, Ecuador
Duración: 22 jun. 202224 jun. 2022

Serie de la publicación

NombreLecture Notes in Networks and Systems
Volumen607 LNNS

Conferencia

Conferencia8th International Conference on Science, Technology and Innovation for Society, CITIS 2022
País/TerritorioEcuador
CiudadGuayaquil
Período22/06/2224/06/22

Nota bibliográfica

Publisher Copyright:
© 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Huella

Profundice en los temas de investigación de 'Classification of Mechanical Failures in Provoked Ignition Engine by Means of ANN and SVM'. En conjunto forman una huella única.

Citar esto