İndiksiyon Motorun Mekanik Arıza Teşhisinde Makine Öğrenme Yöntemlerinin Kullanılması

Abstract

Elektrik makinalarında erken arıza tespiti, arızanın büyüyüp hasarı yaymadan önüne geçilmesi açısından oldukça önemlidir. Arızaların büyümeden öngörülüsü, motorun ömrünü artırabildiğinden araştırmacıların ilgi odağı haline gelmiştir. Bu yönde çalışan araştırmacılar endüstriyel düzeyde hızlı, yorumlaması kolay ve işletme açısından uygulanabilirlik olan teknikler üzerine odaklanmıştır. Bu çalışmada indüksiyon motorlarda oluşan kırık rotor çubuğu ve eksenden kaçıklık arızalarının sonuçlarını sunmaktadır. Sağlıklı ve hatalı koşullar için bir indüksiyon motorun sonlu elemanlar modeli (FEM) geliştirilmiş ve analiz edilmiştir. Arızalı bir makinenin modeli, sağlıklı motorun fiziksel durum ve mekanik pozisyonları değiştirilip farklı arıza şiddetleri oluşturularak akım, gerilim, akı ve tork sinyalleri incelenmiştir. Bu farklı arıza şiddetlerine ait elektriksel sinyallerin verdiği tepkiler karşılaştırılmıştır. Elde edilen akım sinyaline ait ham verilere hızlı fourier yöntemi (FFT) uygulanarak işlenmiş veriler elde edilmiştir. Öznitelik çıkarımı olarak kNN, MLP, RT gibi farklı sınıflandırma metotları kullanılmıştır. Kırık rotor çubuğuna ait farklı arıza şiddetleri ile ilgilenirken, eksantriklik arızasında ise statik eksantriklik, dinamik eksantriklik ve karışık eksantriklik arızaları üzerinde durulmuştur. Ayrıca, farklı sınıflandırmalar kullanarak karşılaştırma yapılmıştır. k-NN, MLP ve RF algoritması sınıflandırma da doğruluğunun oldukça belirgin olduğu tespit edilmiştir.

Keywords

• İndiksiyon motor
• veri madenciliği
• öznitelik çıkarımı
• arıza tespiti

The Use of Machine Learning Methods For Induction Motor Mechanical Fault Diagnosis

Abstract

Early fault detection in electrical machines, grow and damage is quite important in terms of preventing the fault from spreading. Predictions of fault from growth have become the focus of attention of researchers as they can increase the life of the motor. Researchers working in this field have focused on techniques that are fast on the industrial level, easy to interpret and applicable to the enterprise. In this study, it presents the results of broken rotor bar and eccentric faults in induction motor. The finite element model (FEM) of an induction motor was developed and analyzed for healthy and defective conditions. The model of a fault machine, the physical state of the healthy motor and the mechanical positions are changed and the current, voltage, flux and torque signals are examined by creating different fault intensities. The responses of electrical signals of these different fault intensities were compared. The processed data were obtained by applying the fast fourier method (FFT) to the raw data of the obtained current signal. As a feature extraction, kNN, MLP, RT with different classification methods are used for training purposes in diagnostics. While dealing with the different fault intensities of the broken rotor bar, static eccentricity, dynamic eccentricity and mixed eccentricity faults are emphasized in the case of eccentricity fault. In addition, comparisons were made using different classifications. The accuracy of k-NN, MLP and RF algorithm classification was found to be quite significant.

Keywords

• Induction motor
• data mining
• feature extraction
• fault detection

References

1. A.Ghoggala, S.E.Zouzoua, H.Razikb, M.Sahraouia, & A.Khezzarc. (May 2009). An improved model of induction motors for diagnosis purposes – Slot skewing effect and air–gap eccentricity faults. ELSEVIER, 50(5), 1336-1347.
2. Adem, T. (2 - 4 Şubat 2011). Veri Madenciliği Süreçleri Veaçık Kaynak Kodlu Veri Madenciliği Araçları. Akademik Bilişim. Malatya.
3. Ceban, A., Pusca, R., & Romary, R. (29 July 2011). Study of Rotor Faults in Induction Motors Using External Magnetic Field Analysis. IEEE, 59(5), 2082 - 2093.Danilo Granda, 1. G., Aguilar, W. G., Arcos-Aviles, D., & Sotomayor, D. (11 April 2017). Broken Bar Diagnosis for Squirrel Cage Induction Motors Using Frequency Analysis Based on MCSA and Continuous Wavelet Transform. MCA, 22(2).
4. Desheng, L., Beibei, Y., Yu, Z., & Jinping, S. (25-27 May 2012). Time-frequency analysis based on BLDC motor fault detection using Hermite S-method. IEEE. Zhangjiajie, China.
5. Dorrell, D., Chindurza, I., & Cossar, C. (17 October 2005). Effects of rotor eccentricity on torque in switched reluctance Machines. IEEE, 41(10), 3961 - 3963.
6. Dorrell, D., Thomson, W., & Roach, S. (Jan/Feb 1997 ). Analysis of airgap flux, current, and vibration signals as a function of the combination of static and dynamic airgap eccentricity in 3-phase induction motors. IEEE, 33(1), 24 - 34.
7. Elif, A., Goşenay, İ., & Mehmet, H. (2014). Asenkron Motorda Statik Eksenden Kaçıklık Arızasının İncelenmesi. Tokat ,TÜRKİYE: Akademik platform.Esfahani, E. T., Wang, S., & Sundararajan, V. ( 17 May 2013). Multisensor Wireless System for Eccentricity and Bearing Fault Detection in Induction Motors. IEEE, 19(3), 818 - 826.
8. Faiz, J., & Sharifian, B. M. (05 May 2006). Different Faults and Their Diagnosis Techniques in Three-Phase Squirrel-Cage Induction Motors—A Review. Electromagnetic, 26(7), 543-569.

9. FeyzaGürbüz, F. ( October 2018). Rule extraction for tram faults via data mining for safe transportation. ELSEVIER, 568-579.
10. Granda, D., Aguilar, W. G., Arcos-Aviles, D., & Sotomayor, D. (2017). Broken Bar Diagnosis for Squirrel Cage Induction Motors Using Frequency Analysis Based on MCSA and Continuous Wavelet Transform. MCA, 22(2).
11. Ishkova, I., & Vítek, O. (20 July 2015). Diagnosis of eccentricity and broken rotor bar related faults of induction motor by means of motor current signature analysis. IEEE. Kouty nad Desnou, Czech Republic.
12. Jee-Hoon, J., Jong-Jae, L., & Bong-Hwan, K. (30 November 2006 ). Online Diagnosis of Induction Motors Using MCSA. IEEE, 53(6), 1842 - 1852.
13. Joksimovic, G., Durovic, M., Penman, J., & Arthur, N. (June 2000). Dynamic simulation of dynamic eccentricity in induction machines-winding function approach. IEEE, 15(2), 143 - 148.Khalid, S., & Galina, M. (10-13 May 2015). Space-time representation of the main air gap flux of a three phase squirrel cage induction motor and its application to detect eccentricity. IEEE. Coeur d'Alene, ID, USA.
14. Mustafa, M., Nikolakopoulos, G., & Gustafsson. (February 2015). Broken bars fault diagnosis based on uncertainty bounds violation for three-phase induction motors. Electrical Energi Systems, 25(2), 304-325.
15. Puche-Panadero, R., Pineda-Sanchez, M., Riera-Guasp, M., Roger-Folch, J., Hurtado-Perez, E., & Perez-Cru, J. (13 January 2009). Improved Resolution of the MCSA Method Via Hilbert Transform, Enabling the Diagnosis of Rotor Asymmetries at Very Low Slip. IEEE, 24(1), 52 - 59.
16. Subramaniam, A., Sahoo, A., Manohar, S. S., & Panda, S. K. (11-14 Aug. 2017). Fault diagnostic techniques for induction machines through finite element analysis. IEEE. Sydney, NSW, Australia.
17. Wiem, Z., Yemna, B., & Hafedh, T. (16-19 March 2015). Co-simulation of induction motor fed by PWM inverter under a broken bar fault. IEEE. Mahdia, Tunisia.Yassa, N., Rachek, M., & Houassine, H. (April 2019). Motor Current Signature Analysis for The Air Gap Eccentricity Detection In The Squirrel Cage Induction Machines. ELSEVIER, 251-262.
18. Yemna Bensalem, H. T. (2015). Analysis of Induction Motor with Stator Winding Short-circuit Fault by Finite Element Model. IJSET, 53(58), 2356-5608.