Gaz Türbinlerinden Kaynaklanan CO ve NOx Emisyonlarının Tahmininde Aykırı Değer İşleme ve Topluluk Regresyon Yaklaşımlarının Kullanımı

Öz

Gaz türbinleri, yüksek verimlilikleri nedeniyle enerji üretim tesislerinde yaygın olarak kullanılmaktadır; ancak, aynı zamanda CO ve NOx gibi zararlı gaz emisyonlarına da neden olmaktadırlar. Bu çalışma, gaz türbinlerinden kaynaklanan CO ve NOx emisyonlarını tahmin etmek için makine öğrenmesi algoritmalarını kullanarak tahmin modelleri geliştirmeye odaklanmaktadır. Kullanılan veri seti, Türkiye'deki bir kombine çevrim gaz türbininden (CCGT) 2011 ve 2015 yılları arasında saatlik olarak toplanan emisyon verilerini içermektedir. Veri setine Z-Skoru, Çeyrekler Arası Aralık (IQR) ve Mahalanobis Mesafesi (MD) gibi çeşitli aykırı değer işleme yöntemleri uygulanarak modellerin performansına etkisine incelenmiştir. Modeller oluşturulurken Rastgele Orman, Ekstra Ağaçlar, Doğrusal Regresyon, Destek Vektör Regresyonu, Karar Ağacı ve K-En Yakın Komşu gibi makine öğrenmesi algoritmaları kullanılmış ve performansları karşılaştırılmıştır. Ayrıca, Gradient Boosting, LightGBM ve CatBoost algoritmalarını temel temel öğrenici ve XGBoost'u meta-öğrenici olarak kullanan Oylama Topluluk Regresyonu (VR) ve İstifleme Topluluk Regresyonu (SR) yöntemlerinin de performansları incelenmiştir. Sonuçlar, IQR yöntemiyle işlenen veri seti üzerinde uygulanan SR modelinin hem NOx hem de CO emisyonları için en yüksek tahmin doğruluğunu sağladığını göstermektedir. Modelin R² değeri NOx için 0.9194, CO için 0.8556 olarak bulunmuş; RMSE ise sırasıyla 2.7669 ve 0.4619 olarak elde edilmiştir. IQR yöntemiyle elde edilen iyileştirilmiş veri kalitesi, SR modelinin üstün performans göstermesine katkı sağlamakta ve modelin NOx ve CO emisyonlarını yüksek hassasiyetle tahmin edebilmesi açısından güvenilir bir yaklaşım olduğunu ortaya koymaktadır.

Anahtar Kelimeler

• Gaz türbini emisyonları
• Makine öğrenmesi
• Aykırı değer işleme
• Kombine çevrim enerji üretimi
• Çeyrekler arası aralık
• Mahalanobis uzaklığı

Outliers Treatment for Improved Prediction of CO and NOx Emissions from Gas Turbines Using Ensemble Regressor Approaches

Öz

Gas turbines are widely used in power generation plants due to their high efficiency, but they also emit pollutants such as CO and NOx. This study focuses on developing predictive models for predicting CO and NOx emissions from gas turbines using machine learning algorithms. The dataset used includes pollutant emission data from a combined cycle gas turbine (CCGT) in Türkiye, collected hourly between 2011 and 2015. Various outlier treatment methods such as Z-Score, Interquartile Range (IQR), and Mahalanobis Distance (MD) are applied to the dataset. Machine learning algorithms including Random Forest, Extra Trees, Linear Regression, Support Vector Regression, Decision Tree, and K-Nearest Neighbors are used to build the predictive models, and their performances are compared. Additionally, Voting Ensemble Regressor (VR) and Stacking Ensemble Regressor (SR) methods are employed, using Gradient Boosting, LightGBM, and CatBoost as base learners and XGBoost as a meta-learner. The results demonstrate that the SR model, when applied to the dataset processed using the IQR method, achieves the highest prediction accuracy for both NOx and CO emissions, with R² values of 0.9194 and 0.8556, and RMSE values of 2.7669 and 0.4619, respectively. These findings highlight the significant role of the IQR method in enhancing model accuracy by effectively handling outliers and reducing data noise. The improved data quality achieved through this method contributes to the superior performance of the SR model, making it a reliable approach for predicting NOx and CO emissions with high precision.

Anahtar Kelimeler

• Gas turbine emissions
• Machine learning
• Outlier processing
• Combined cycle power generation
• Interquartile range
• Mahalanobis distance

Kaynakça

1. Ahmad, M. W., Reynolds, J., & Rezgui, Y. (2018). Predictive modelling for solar thermal energy systems: A comparison of support vector regression, random forest, extra trees and regression trees. Journal of Cleaner Production, 203, 810-821. https://doi.org/10.1016/j.jclepro.2018.08.207
2. Aslan, E. (2024). Prediction and Comparative Analysis of Emissions from Gas Turbines Using Random Search Optimization and Different MachineL earning Based Algorithms. Bulletin of the Polish Academy of Sciences Technical Sciences, e151956-e151956. https://doi.org/ 10.24425/bpasts.2024.151956
3. Biau, G. (2012). Analysis of a random forests model. The Journal of Machine Learning Research, 13(1), 1063-1095.
4. Caicedo, J. C., Cooper, S., Heigwer, F., Warchal, S., Qiu, P., Molnar, C., ... & Carpenter, A. E. (2017). Data-analysis strategies for image-based cell profiling. Nature Methods, 14(9), 849-863.
5. Coelho, D. S. L., Ayala, H. V. H., & Mariani, V. C. (2024). CO and NOx emissions prediction in gas turbine using a novel modeling pipeline based on the combination of deep forest regressor and feature engineering. Fuel, 355, 129366. https://doi.org/10.1016/j.fuel.2023.129366
6. Dalal, A. S., Sultanova, N., Jayabalan, M., & Mustafina, J. (2023, December). Gas turbine–CO & NOx emission data analysis with predictive modelling using ML/AI approaches. In 2023 16th International Conference on Developments in eSystems Engineering (DeSE) (pp. 100-104). IEEE. https://doi.org/10.1109/DeSE60595.2023.10469322
7. Dirik, M. (2022). Prediction of NOx emissions from gas turbines of a combined cycle power plant using an ANFIS model optimized by GA. Fuel, 321, 124037. https://doi.org/10.1016/j.fuel.2022.124037
8. Divina, F., Gilson, A., Goméz-Vela, F., García Torres, M., & Torres, J. F. (2018). Stacking ensemble learning for short-term electricity consumption forecasting. Energies, 11(4), 949. https://doi.org/10.3390/en11040949

9. Farzaneh-Gord, M., & Deymi-Dashtebayaz, M. (2011). Effect of various inlet air cooling methods on gas turbine performance. Energy, 36(2), 1196-1205. https://doi.org/10.1016/j.energy.2010.11.027
10. Ghorbani, H. (2019). Mahalanobis distance and its application for detecting multivariate outliers. Facta Universitatis, Series: Mathematics and Informatics, 583-595. https://doi.org/10.22190/FUMI1903583G
11. Karthikeyan, S., Kathirvalavakumar, T., & Prasath, R. (2023, June). Classification of the Class Imbalanced Data Using Mahalanobis Distance with Feature Filtering. In International Conference on Mining Intelligence and Knowledge Exploration (pp. 45-53). Cham: Springer Nature Switzerland.
12. Kaya, H., Tüfekci, P., & Uzun, E. (2019). Predicting CO and NOx emissions from gas turbines: novel data and a benchmark PEMS. Turkish Journal of Electrical Engineering and Computer Sciences, 27(6), 4783-4796. https://doi.org/10.3906/elk-1807-87
13. Kochueva, O., & Nikolskii, K. (2021). Data analysis and symbolic regression models for predicting CO and NOx emissions from gas turbines. Computation, 9(12), 139. https://doi.org/10.3390/computation9120139
14. Kumar, M. V., Babu, A. V., Reddy, C. R., Pandian, A., Bajaj, M., Zawbaa, H. M., & Kamel, S. (2022). Investigation of the combustion of exhaust gas recirculation in diesel engines with a particulate filter and selective catalytic reactor technologies for environmental gas reduction. Case Studies in Thermal Engineering, 40, 102557.
15. Leys, C., Klein, O., Dominicy, Y., & Ley, C. (2018). Detecting multivariate outliers: Use a robust variant of the Mahalanobis distance. Journal of Experimental Social Psychology, 74, 150-156. https://doi.org/10.1016/j.jesp.2017.09.011
16. Liu, Z., Meng, H., Huang, J., Kwangwari, P., Ma, K., Xiao, B., & Li, L. (2021). Acute carbon monoxide poisoning with low saturation of carboxyhaemoglobin: a forensic retrospective study in Shanghai, China. Scientific Reports, 11(1), 18554.
17. Lopes, C., Antelo, L. T., Franco-Uría, A., Alonso, A. A., & Pérez-Martín, R. (2015). Valorisation of fish by-products against waste management treatments–Comparison of environmental impacts. Waste Management, 46, 103-112. https://doi.org/10.1016/j.wasman.2015.08.017
18. Lott, P., Casapu, M., Grunwaldt, J. D., & Deutschmann, O. (2024). A review on exhaust gas after-treatment of lean-burn natural gas engines–From fundamentals to application. Applied Catalysis B: Environmental, 340, 123241.
19. Mahmoudi, S., Baeyens, J., & Seville, J. P. (2010). NOx formation and selective non-catalytic reduction (SNCR) in a fluidized bed combustor of biomass. Biomass and bioenergy, 34(9), 1393-1409. https://doi.org/10.1016/j.biombioe.2010.04.013
20. Mare, D. S., Moreira, F., & Rossi, R. (2017). Nonstationary Z-score measures. European Journal of Operational Research, 260(1), 348-358. https://doi.org/10.1016/j.ejor.2016.12.001
21. Maulud, D., & Abdulazeez, A. M. (2020). A review on linear regression comprehensive in machine learning. Journal of Applied Science and Technology Trends, 1(2), 140-147. https://doi.org/10.38094/jastt1457
22. Mishra, P., Pandey, C. M., Singh, U., Gupta, A., Sahu, C., & Keshri, A. (2019). Descriptive statistics and normality tests for statistical data. Annals of Cardiac Anaesthesia, 22(1), 67-72. https://doi.org/10.4103/aca.ACA_157_18
23. Naghibi, A. (2024). Exploring explainable ensemble machine learning methods for long-term performance prediction of industrial gas turbines: A comparative analysis. Engineering Applications of Artificial Intelligence, 138, 109318. https://doi.org/10.1016/j.engappai.2024.109318
24. Nino-Adan, I., Portillo, E., Landa-Torres, I., & Manjarres, D. (2021). Normalization influence on ANN-based models performance: A new proposal for Features’ contribution analysis. IEEE Access, 9, 125462-125477. https://doi.org/10.1109/ACCESS.2021.3110647
25. Osborne, J. W., & Overbay, A. (2019). The power of outliers (and why researchers should always check for them). Practical Assessment, Research, and Evaluation, 9(1), 6. https://doi.org/10.7275/qf69-7k43
26. Pachauri, N. (2024). An emission predictive system for CO and NOx from gas turbine based on ensemble machine learning approach. Fuel, 366, 131421. https://doi.org/10.1016/j.fuel.2024.131421
27. Pandey, R. A., & Chandrashekhar, B. (2014). Physicochemical and biochemical approaches for treatment of gaseous emissions containing NOx. Critical Reviews in Environmental Science and Technology, 44(1), 34-96. https://doi.org/10.1080/10643389.2012.710430
28. Prusty, S., Patnaik, S., & Dash, S. K. (2022). SKCV: Stratified K-fold cross-validation on ML classifiers for predicting cervical cancer. Frontiers in Nanotechnology, 4, 972421. https://doi.org/10.3389/fnano.2022.972421
29. Quinlan, J. R. (1996). Learning decision tree classifiers. ACM Computing Surveys (CSUR), 28(1), 71-72.
30. Rezazadeh, A. (2021). Environmental pollution prediction of NOx by predictive modelling and process analysis in natural gas turbine power plants. Pollution, 7(2), 481-494. https://doi.org/10.22059/poll.2021.316327.977
31. Song, Y., Liang, J., Lu, J., & Zhao, X. (2017). An efficient instance selection algorithm for k nearest neighbor regression. Neurocomputing, 251, 26-34. https://doi.org/10.1016/j.neucom.2017.04.018
32. Tian, J., Wang, L., Xiong, Y., Wang, Y., Yin, W., Tian, G., ... & Ji, S. (2024). Enhancing combustion efficiency and reducing nitrogen oxide emissions from ammonia combustion: A comprehensive review. Process Safety and Environmental Protection, 183, 514-543. https://doi.org/10.1016/j.psep.2024.01.020
33. Todeschini, R., Ballabio, D., Consonni, V., Sahigara, F., & Filzmoser, P. (2013). Locally centred Mahalanobis distance: a new distance measure with salient features towards outlier detection. Analytica Chimica Acta, 787, 1-9. https://doi.org/10.1016/j.aca.2013.04.034
34. Valkenborg, D., Rousseau, A. J., Geubbelmans, M., & Burzykowski, T. (2023). Support vector machines. American Journal of Orthodontics and Dentofacial Orthopedics, 164(5), 754-757.
35. Wardana, M. K. A., & Lim, O. (2022). Review of improving the NOx conversion efficiency in various diesel engines fitted with SCR system technology. Catalysts, 13(1), 67.
36. Wood, D. A. (2023). Long-term atmospheric pollutant emissions from a combined cycle gas turbine: Trend monitoring and prediction applying machine learning. Fuel, 343, 127722. https://doi.org/10.1016/j.fuel.2023.127722
37. Wu, T. J., Burke, J. P., & Davison, D. B. (1997). A measure of DNA sequence dissimilarity based on Mahalanobis distance between frequencies of words. Biometrics, 1431-1439. https://doi.org/10.2307/2533509
38. Yang, L., & Shami, A. (2020). On hyperparameter optimization of machine learning algorithms: Theory and practice. Neurocomputing, 415, 295-316. https://doi.org/10.1016/j.neucom.2020.07.061
39. Yaro, A. S., Maly, F., & Prazak, P. (2023). Outlier detection in time-series receive signal strength observation using z-score method with s n scale estimator for indoor localization. Applied Sciences, 13(6), 3900. https://doi.org/10.3390/app13063900
40. Yousif, S. T., Ismail, F. B., & Al‐Bazi, A. (2024). A hybrid neural network‐based improved PSO algorithm for gas turbine emissions prediction. Advanced Theory and Simulations, 2301222.
41. https://doi.org/10.1002/adts.202301222 Yu, H., & Kim, S. (2012). SVM Tutorial-Classification, Regression and Ranking. Handbook of Natural Computing, 1, 479-506.