Agosto 2024 vol. 11 num. 1 - XXXI Simpósio Internacional de Engenharia Automotiva

Trabalho completo - Open Access.

Idioma principal | Segundo idioma

Clusterização não supervisionada para monitoramento da saúde de motores de combustão interna

Unsupervised Clustering for Internal Combustion Engines Health Monitoring

JUNQUEIRA, B. F. ; VIEIRA, R. G. ; PINTO, I. J. P. ; KARAZIACK, B. B. ; FREITAS, I. J. V. ;

Trabalho completo:

É apresentada uma abordagem de aprendizado de máquina não supervisionada destinada a auxiliar o monitoramento da saúde de motores de combustão interna, possibilitando a detecção precoce de possíveis falhas ou degradação ao longo do tempo. Nossa metodologia utiliza dados temporais coletados de um dispositivo datalogger conectado ao sistema On-Board Diagnostics de um veículo. Esses dados servem como entrada para um modelo de aprendizado de máquina de agrupamento, treinado incrementalmente ao longo do tempo para detectar anomalias em séries temporais multivariadas. O processo de tomada de decisão para categorizar comportamentos de clusters como anomalias, que podem ser indícios do começo da degradação do motor ao longo do tempo, baseia-se em diversas métricas-chave, como o coeficiente de similaridade de Jaccard, população relativa do cluster, estabilidade e movimento do centróide ao longo do tempo. A abordagem proposta é validada em um conjunto de dados públicos de uma máquina rotativa e testada em um motor de combustão interna de um veículo de tamanho médio em condição de ponto morto.

Trabalho completo:

We present an unsupervised machine learning approach designed to aid in the health monitoring of internal combustion engines, enabling the early detection of possible failures or degradation over time. Our methodology uses temporal data collected from a datalogger device connected to a vehicle

Download (PDF)

Palavras-chave: -,

Palavras-chave: -,

DOI: 10.5151/simea2024-PAP32

Referências bibliográficas
  • [1] " Henriquez P, Alonso J B, Ferrer M A, Travieso C
  • [2] M. Review of Automatic Fault Diagnosis Systems Using
  • [3] Audio and Vibration Signals. IEEE Transactions on
  • [4] Systems, Man and Cybernetics: Systems, 2013.
  • [5] [2] Zhang Q, Zuo Z, Liu J. Failure analysis of a diesel
  • [6] engine cylinder head based on finite element method,
  • [7] Engineering Failure Analysis, vol 34, 2013.
  • [8] [3] Liu X F, Wang Y, Liu W H. Finite element
  • [9] analysis of thermo-mechanical conditions inside the piston
  • [10] of a diesel engine, Applied Thermal Engineering, vol 119,
  • [11] 2017.
  • [12] [4] Ghorpade U S, Chavan D S, Patil V, Gaikwad M.
  • [13] Finite element analysis and natural frequency optimization
  • [14] of engine bracket, International Journal of Mechanical and
  • [15] Industrial Engineering, vol 3, 2013.
  • [16] [5] Seralathan S, Mitnala S V, Reddy RV S K, Venkat
  • [17] I G, Reddy D R T, Hariram V, Premkumar T M. Stress
  • [18] analysis of the connecting rod of compression ignition
  • [19] engine, Materials Today: Proceedings, vol 33, 2020.
  • [20] [6] Avwunuketa A, Enyia J, Oloruntoba S. Numerical
  • [21] and Thermal Finite Element Analysis (FEA) of Idealized
  • [22] Gas Turbine Engine Blade, International Journal of
  • [23] Aerospace and Mechanical Engineering, vol 7, 2020.
  • [24] [7] Sun L, Shang Z, Bhowmick S, Nagarajaiah S.
  • [25] Review of Bridge Structural Health Monitoring Aided by
  • [26] Big Data and Artificial Intelligence: From Condition
  • [27] Assessment to Damage Detection, Journal of Structural
  • [28] Engineering, vol 146, 2020.
  • [29] [8] Zhao R, Yan R, Chen Z, Wang P, Gao R X. Deep
  • [30] learning and its applications to machine health monitoring.
  • [31] Mechanical Systems and Signal Processing, 2019.
  • [32] [9] Dong S, He K, Tang B. The fault diagnosis
  • [33] method of rolling bearing under variable working
  • [34] conditions based on deep transfer learning. Journal of the
  • [35] Brazilian Society of Mechanical Sciences and Engineering,
  • [36] vol 42, 2020.
  • [37] [10] Zhang Q, Lin J, Song H, Sheng G. Fault
  • [38] Identification Based on PD Ultrasonic Signal Using RNN,
  • [39] DNN and CNN. Condition Monitoring and Diagnosis
  • [40] (CMD), 2018.
  • [41] [11] Liang P, Deng C, Wu J, Yang Z, Zhu J. Intelligent
  • [42] Fault Diagnosis of Rolling Element Bearing Based on
  • [43] Convolutional Neural Network and Frequency
  • [44] Spectrograms. IEEE International Conference on
  • [45] Prognostics and Health Management, 2019.
  • [46] [12] Ramteke S M, Chelladurai H, Amarnath M.
  • [47] Diagnosis and Classification of Diesel Engine Components
  • [48] Faults Using Time-Frequency and Machine Learning
  • [49] Approach. Journal of Vibration Engineering &
  • [50] Technologies, 2021.
  • [51] [13] Carrera F, Falchi F, Girardi M, Messina N,
  • [52] Padovani C, Pellegrini D. Deep learning for structural
  • [53] health monitoring: An application to heritage structures.
  • [54] Cornell University Electrical Engineering and Systems
  • [55] Science, 2022.
  • [56] [14] Junqueira B F, Leiderman R, Castello D A.
  • [57] Damage recovery in composite laminates through deep
  • [58] learning from acoustic scattering of guided waves.
  • [59] Ultrasonics, vol 139, 2024.
  • [60] [15] Xiong M, Wang H, Fu Q. Digital twin–driven
  • [61] aero-engine intelligent predictive maintenance. The
  • [62] International Journal of Advanced Manufacturing
  • [63] Technology, vol 114, 2021
  • [64] [16] Jan S U, Lee Y D, Koo I S. A distributed
  • [65] sensor-fault detection and diagnosis framework using
  • [66] machine learning. Information Sciences, vol 547, 2021.
  • [67] [17] Chukwudi I J, Zaman N, Rahim M A, Rahman M
  • [68] A, Alenazi M J F, Pillai P. An Ensemble Deep Learning
  • [69] Model for Vehicular Engine Health Prediction. IEEE
  • [70] Access, vol 12, 2024.
  • [71] [18] Liao J, Hu J, Yan F, Chen P, Zhu L, Zhou Q, Xu
  • [72] H, Li J. A comparative investigation of advanced machine
  • [73] learning methods for predicting transient emission
  • [74] characteristic of diesel engine. Fuel, vol 350, 2023.
  • [75] [19] Khurana S, Saxeng S, Jain S, Dixit A. Predictive
  • [76] modeling of engine emissions using machine learning: A
  • [77] review. Materials Today: Proceedings, vol 38, 2021.
  • [78] [20] Xu X, Zhao Z, Xu X, Yang J, Chang L, Yan X,
  • [79] Wang G. Machine learning-based wear fault diagnosis for
  • [80] marine diesel engine by fusing multiple data-driven models.
  • [81] Knowledge-Based Systems, vol 190, 2020.
  • [82] [21] Tang M, Chen H, Guan C. Research on diesel
  • [83] engine fault diagnosis method based on machine learning.
  • [84] 4th International Conference on Frontiers Technology of
  • [85] Information and Computer (ICFTIC), 2022.
  • [86] [22] Ramteke S M, Chelladurai H, Amarnath M.
  • [87] Diagnosis and Classification of Diesel Engine Components
  • [88] Faults Using Time–Frequency and Machine Learning
  • [89] Approach. Journal of Vibration Engineering &
  • [90] Technologies, vol 10, 2021.
  • [91] [23] Junqueira B F, Violato R P V, Simões F O,
  • [92] Tuleski B. Aplicação de Machine Learning para
  • [93] Diagnóstico de Falha em Motores a Diesel utilizando Sinais
  • [94] de Áudio. Anais do XXIX Simpósio Internacional de
  • [95] Engenharia Automotiva, 2022.
  • [96] [24] Aliramezani M, Koch C R, Shahbakhti M.
  • [97] Modeling, diagnostics, optimization, and control of internal
  • [98] combustion engines via modern machine learning
  • [99] techniques: A review and future directions. Progress in
  • [100] Energy and Combustion Science, vol 88, 2022.
  • [101] [25] Landauer M, Wurzenberger M, Skopik F,
  • [102] Settanni G, Filzmoser P. Time series analysis: Unsupervised
  • [103] anomaly detection beyond outlier detection. In C. Su & H.
  • [104] Kikuch (Eds.), Information Security Practice and
  • [105] Experience, 2018.
  • [106] [26] Niwattanakul S, Singthongchai J, Naenudorn E,
  • [107] Wanapu S. Using of Jaccard Coefficient for Keywords
  • [108] Similarity. Proceedings of the International
  • [109] MultiConference of Engineers and Computer Sc, vol 1,
  • [110] 2013.
  • [111] [27] Greene D, Doyle D, Cunningham P. Tracking the
  • [112] evolution of communities in dynamic social networks.
  • [113] International Conference on Advances in Social Network
  • [114] Analysis and Mining, 2010.
  • [115] [28] Toyoda M, Kitsuregawa M. Extracting Evolution
  • [116] of Web Communities from a Series of Web Archives.
  • [117] Association for Computing Machinery, 2003.
  • [118] [29] Zhou A, Cao F, Qian W, Jin C. Tracking clusters
  • [119] in evolving data streams over sliding windows. Knowledge
  • [120] and Information Systems, vol 15, 2008.
  • [121] [30] Jung W, Kim S H, Yun S H, Bae J, Park Y H.
  • [122] Vibration, acoustic, temperature, and motor current dataset
  • [123] of rotating machine under varying operating conditions for
  • [124] fault diagnosis. Data in Brief, vol 48, 2023.
  • [125] [31] Lee J, Kim J, Bae C. Effect of the
  • [126] air-conditioning system on the fuel economy in a gasoline
  • [127] engine vehicle. Proceedings of the Institution of
  • [128] Mechanical Engineers, 2013."
Como citar:

JUNQUEIRA, B. F.; VIEIRA, R. G.; PINTO, I. J. P.; KARAZIACK, B. B.; FREITAS, I. J. V.; "Clusterização não supervisionada para monitoramento da saúde de motores de combustão interna", p. 192-201 . In: Anais do XXXI Simpósio Internacional de Engenharia Automotiva . São Paulo: Blucher, 2024.
ISSN 2357-7592, DOI 10.5151/simea2024-PAP32

últimos 30 dias | último ano | desde a publicação


downloads


visualizações


indexações