Title: Detection and Classification of Ring Failures by Artificial Neural Networks in Bearings

Abstract: An effective way to improve the efficiency and extend the life of the machines is to determine the failures of the bearings during operation. Early detection of bearing failures also has critical importance in terms of production costs. Various maintenance methods are used to prevent the failures. Despite all the precautions, unexpected failures can occur and production operations can be failed. This situation, apart from conventional methods, requires a novel determination and diagnostic technique. In this study, artificial intelligence based methods are applied and models are developed in order to detect bearing failures early and to classify the type of failure. In the developed models, it is possible to detect the ring failures depending on different loads and bearing vibration information. In addition, a classification is carried out for the fault from the inner or outer race of the bearing. Determination of the fault, as well as the diagnosis of the class, will increase stability and productivity, especially in critical industrial applications.

Keywords: Bearing; ring failures; classification

Başlık: Rulmanlarda Bilezik Arızalarının Yapay Sinir Ağları ile Tespiti ve Sınıflandırılması

Özet: Makinelerin verimini arttırmanın ve ömrünü uzatmanın etkili bir yolu rulmanların çalışma sırasında arızalarının tespit edilmesidir. Rulman arızalarının erken tespit edilmesi üretim maliyetleri açısından da kritik öneme sahiptir. Arızaları önlemek için çeşitli bakım yöntemleri kullanılmaktadır. Tüm önlemlere rağmen beklenmedik arızalar oluşabilmekte ve üretim faaliyetleri aksayabilmektedir. Bu durum geleneksel yöntemlerin dışında sıra dışı bir arıza tespit ve teşhis tekniğine ihtiyacı doğurmuştur. Bu çalışmada rulman arızalarını erkenden tespit etmek ve arıza tipini sınıflandırmak amacıyla yapay zekâ tabanlı yöntemler uygulanmakta ve modeller geliştirilmektedir. Geliştirilen modellerde farklı yüklenmelere ve rulman titreşim bilgisine bağlı olarak bilezik hatası tespiti yapılmaktadır. Ayrıca hatanın iç ya da dış bilezikten kaynaklandığına yönelik bir sınıflandırma gerçekleştirilmektedir. Gerek hatanın tespiti gerekse de sınıfının belirlenmesi özellikle kritik endüstriyel uygulamalarda kararlı çalışmayı ve verimliliği artıracaktır.

Anahtar kelimeler: Rulman; bilezik hatası; sınıflandırma

• Kalyoncu M. Titresim analizi ile makina elemanlari arizalarinin belirlenmesi. Muhendis ve Makina Dergisi 2006; 47(552): 28-35.
• Orhan S, Arslan H, Akturk N. Determination of rolling element bearing defects through vibration analysis. Journal of the Faculty of Engineering and Architecture of Gazi University 2003; 18(2): 39-48.
• Aliustaoglu C. The real time analysis and diagnosis of bearing faults. Kocaeli University, Fen Bilimleri Enstitusu, MSc Thesis, Kocaeli, Turkey, 2008.
• Unal M. Diagnosis of bearing faults by signal process and genetic neural network. Marmara University, Fen Bilimleri Enstitusu, PhD Thesis, Istanbul, Turkey, 2014.
• Rao BKN, Pai PS, Nagabhushana TN. Failure diagnosis and prognosis of rolling – element bearings using artificial neural networks: A critical overview. Journal of Physics: Conference Series 2012; 364: 1-29.
• Shridhar KU, Pradeep S. Prognosis of rotating machinery using artificial neural network. International Journal of Engineering Research & Technology (IJERT) 2013; 2(8): 1433-1439.
• Jamadar IM, Vakhaira DP. A novel approach integrating dimensional analysis and neural networks for the detection of localized faults in roller bearings. Measurement 2016; 94: 177–185.
• Guo L, Li N, Jia F, Lei Y, Kin J. A recurrent neural network based health indicator for remaining useful life prediction of bearings. Neurocomputing 2017; 240: 98–109.
• Zurita G, Sanchez V, Cabrera D. A reviev vibration machine diagnostics by using artificial intelligence methods. Investigacion & Desarollo 2016; 16(1): 102-114.
• Patel JP, Upadhyay SH. Comparison between artificial neural network and support vector method for a fault diagnostics in rolling element bearings. Procedia Engineering 2016; 144: 390-397.
• Chen X, Zhaou J, Xiao J, Zhang X, Xiao H, Zhu, W, Fu W. Fault diagnosis based on dependent feature vector and probability neural network for rolling element bearings. Applied Mathematics and Computation 2014; 247: 835–847.
• Bechhoefer E. A quick introduction to bearing envelope analysis. Gren Power Monitoring Systems, Cornwall, VT 05752, 2012.
• Acousitcs data. Retrieved from http://data-acoustics.com/measurements/bearing-faults/bearing-2/ at May 16, 2017.
• Kahraman HT. A novel and powerful hybrid classifier method: Development and testing of heuristic k-nn algorithm with fuzzy distance metric. Data & Knowledge Engineering 2016; 103: 44-59.
• Rahman G, Islam Z. Missing value imputation using decision trees and decision forests by splitting and merging records: Two novel techniques. Knowledge-Based Systems 2013; 53: 51–65.
• Nayak J, Naik B, Behera HS. A comprehensive survey on support vector machine in data mining tasks: Applications & challenge. International Journal of Database Theory and Application 2015; 8(1): 169-186.
• Colak I, Bal G, Demirtas M, Kahraman HT. A parameter determination system for wind turbines based on naive Bayes classification algorithm. In The Eighth International Conference on Machine Learning and Applications (ICMLA 2009), Florida, USA, 2009, pp. 611-616.
• Neto A, Bonini C, Bisi B, Coletta L, Reis A. Artificial neural network for classification and analysis of degraded soils. IEEE Latin America Transactions 2017; 15(3): 503-509.