Emre ÇANCIOĞLU, Savas SAHİN, Yalçın İŞLER

Poincare Çizimi Ölçümlerinden Topluluk Öğrenmesi Yöntemleri Kullanılarak Proses Kontrol Sistemlerinde Arıza Tespit ve Teşhisi

Bu çalışmada, farklı kimyasal birimlere ait doğrusal olmayan süreçler içeren bir endüstriyel tesisteki 20 farklı arızanın tespiti ve sınıflandırılması yapılmıştır. Kullanılan veri seti büyük bir endüstriyel tesisten elde edilen IEEEDataPort çevrimiçi veri kümesidir. Tennessee Eastman Süreci olarak bilinen bu veri seti 20 farklı hata türü ile 52 işlem noktasından alınan ölçümleri içerir. Bu ölçümler üzerinden Poincare çizimleri elde edilerek her işlem noktası için sık kullanılan doğrusal olmayan öznitelikler çıkarılmıştır. Bu öznitelikler %5 istatistiksel anlamlılık düzeyinde tek yönlü ANOVA testine uygulanarak hata türleri arasında istatistiksel olarak anlamlı fark olduğunu gösterenler seçilmiştir. Hem tüm öznitelikler hem de sadece ANOVA ile seçilen öznitelikler beş farklı topluluk öğrenmesi algoritması (Boosted Trees, Bagged Trees, Subspace Discriminant, Subspace KNN ve RUSBoosted Trees) kullanılarak sınıflandırılmıştır. Bu çalışmada elde edilen en yüksek sınıflandırıcı doğruluğu Subspace Discriminant algor itması kulanılarak %89,5 olarak elde edilmiştir. Aynı verisetini kullanan benzer çalışmalarla kıyaslanabilir bir başarı düzeyine ulaşılmıştır. Öte yandan, ANOVA tabanlı öznitelik seçiminin bu tür endüstriyel proses tesislerinde arızaların teşhisinde bariz bir üstünlük sağlamadığı görülmüştür.

Anahtar Kelimeler:

Tennessee Eastman Proses Sistemi, Arıza Tespiti, Arıza Teşhisi, Topluluk Öğrenmesi, Poincare Grafik Ölçümleri, Tek Yönlü ANOVA Testi.

Fault Detection and Diagnosis on Process Control Systems Using Ensemble Learning Algorithms from Poincare Plot Measures

This study aimed to detect and classify 20 different malfunctions in an industrial facility that involves nonlinear processes from various chemical units. The IEEEDataPort online dataset, acquired from a large industrial plant, was used in this study. It contains measures from 52 process points in Tennessee Eastman Process with 20 different fault types. We extracted two commonly used nonlinear features from Poincare Plots for each measurement point. The statistically meaningful features, which show statistically significant differences among fault types with a significance of 5%, were selected from these features. Five distinct Ensemble Learner algorithms (Boosted Trees, Bagged Trees, Subspace Discriminant, Subspace KNN, and RUSBoosted Trees) discriminated the fault types using all features and the selected features only. The maximum classifier accuracies were 89.5% for both feature sets using the Subspace Discriminant method in this study. This performance is a comprehendible result among the results achieved in similar studies. On the other hand, ANOVA-based feature selection didn't result in a clear advantage to diagnose faults in such industrial process plants.

Keywords:

Tennessee Eastman Process System, Fault Detection, Fault Diagnosis, Ensemble Learning, Poincare Plot Measures, One Way ANOVA Test.,

PDF

___

Chadha, G. S., & Schwung, A. (2017, September). Comparison of deep neural network architectures for faultdetection in Tennessee Eastman process. In 2017 22nd IEEE International Conference on Emerging Technologies and Factory Automation (ETFA) (pp. 1 8). IEEE.
Hajihosseini, P., Anzehaee, M. M., & Behnam, B. (2018). Fault detection and isolation in the challenging T ennessee Eastman process by using image processing techniques. ISA transactions, 79, 137 146.
D’Angelo, M. F., Palhares, R. M., Camargos Filho, M. C., Maia, R. D., Mendes, J. B., & Ekel, P. Y. (2016). A new fault classification approach applied to Tenn essee Eastman benchmark process. Applied Soft Computing, 49, 676 686.
Downs, J. J., & Vogel, E. F. (1993). A plant wide industrial process control problem. Computers & chemical engineering, 17(3), 245 255.
Ricker, N. L. (1996). Decentralized contro l of the Tennessee Eastman challenge process. Journal of Process Control, 6(4), 205 221.
Nashalji, M. N., Shoorehdeli, M. A., & Teshnehlab, M. (2010). Fault detection of the Tennessee Eastman process using improved PCA and neural classifier. In Soft co mputing in industrial applications (pp. 41 50). Springer, Berlin, Heidelberg.
Puurula, A., Read, J., & Bifet, A. (2014). Kaggle LSHTC4 winning solution. arXiv preprint arXiv:1405.0546.
Niculescu Mizil, A., Perlich, C., Swirszcz, G., Sindhwani, V., Liu, Y., Melville, P., ... & Zhu, Y. F. (2009, December). Winning the KDD cup orange challenge with ensemble selection. In KDD Cup 2009 Competition (pp. 23 34). PMLR.
Schclar, A., Tsikinovsky, A., Rokach, L., Meisels, A., & Antwarg, L. (2009, October). Ensemble methods for improving the performance of neighborhood based collaborative filtering. In Proceedings of the third ACM conference on Recommender systems (pp. 261 264).
Cancioglu, E., Sahin, S., & Isler, Y. (2021). Fault Detection and Diagnosis on Process Control Systems Using k Nearest Neighbors from Poincare Plot Measures. In International Conference on Applied Sciences, Engineering and Mathematics (ICASEM’2021), ACCEPTED, Skopje, North
Mireles Gonzalez, J. I. ( Deep recu rrent neural networks for fault detection and classification (M.Sc. Thesis, University of Waterloo).
Duda, R. O., Hart, P. E. & Stork, D. G. (2001). Pattern Classification. New York: John Wiley and Sons, 2nd Edition.
Isler, Y., Narin, A. & Ozer, M. (2015). Comparison of the effects of cross validation methods on determining performances of classifiers used in diagnosing congestive heart failure. Measurement Science Review, 15(4), 196 201.
Narin, A., Isler, Y. & Ozer, M. (2014). Investigating the performance improvement of HRV Indices in CHF using feature selection methods based on backward elimination and statistical significance. Computers in Biology and Medicine, 45, 72 79.
Akgul, A. (2003). Tıbbi Araştırmalarda İstatistiksel Analiz Tek nikleri: SPSS Uygulamaları. Seçkin Yayıncılık, Ankara, Turkey.
Isler, Y. & Kuntalp, M. (2007). Combining classical HRV indices with wavelet entropy measures improves to performance in diagnosing congestive heart failure. Computers in Biology and Medic ine, 37(10), 1502 1510.
Isler, Y. (2009). A Detailed Analysis of the Effects of Various Combinations of Heart Rate Variability Indices in Congestive Heart Failure. PhD thesis, Department of Electrical and Electronics Engineering, The Graduate School o f Natural and Applied Sciences, Dokuz Eylul University.
Marciano, M. L., Migaux, F., Acanfora, D., Furgi, G. & Rengo, R. (1994). Quantification of poincare maps for the evaluation of heart rate variability. Computers in Caridology, 1994, 557 580. [ Brennan, M., Palaniswami, M. & Kamen, P. (2001). Do existing measures of poincare plot geometry reflect nonlinear features of heart rate variability? IEEE Transactions on Biomedical Engineering, 48(11), 1342 1347.
Chen, X. (2019). Tennessee Eastman s imulation dataset. IEEE Dataport, https://dx.doi.org/10.21227/4519 z502.