Makine Öğrenmesi Modelleri ile Lamine Ahşap Kirişlerin Eğilme Dayanımı (MOR) Değerlerinin Tahmini

Abstract

Bu çalışma, farklı geometrilere ve üretim parametrelerine sahip yapıştırılmış lamine ahşap (glulam) kirişlerin eğilme dayanımı (MOR) değerlerinin üç nokta (3P) ve dört nokta (4P) eğilme testleri üzerinden belirlenmesini ve bu değerlerin makine öğrenmesi tabanlı modeller kullanılarak tahmin edilmesini amaçlamaktadır. Deneysel kirişler, PVAc-D3 (Polivinil Asetat) tutkallı sarıçam (Pinus sylvestris L.) lameller kullanılarak düz, eğrisel, kontrplak takviyeli ve takviyesiz olmak üzere dört farklı grupta üretilmiştir. Elde edilen deneysel veriler, beş farklı regresyon algoritması (SVR, Lasso, Ridge, Huber ve XGBoost) ile analiz edilmiş ve her modelin tahmin performansı 10 katlı çapraz doğrulama yöntemiyle değerlendirilmiştir. Analiz sonuçlarına göre SVR modeli, üç noktalı eğilme deneylerinde R² = 0,9920 ve RMSE = 0,4587 MPa gibi oldukça yüksek doğruluk ve düşük hata değerleriyle en başarılı performansı göstermiştir. Dört noktalı deneylerinde de benzer bir eğilim gözlenmiş, SVR modeli R² = 0,9782 ile üstün performansını korumuştur. Bu bulgular, makine öğrenmesi tabanlı yaklaşımların, glulam kirişlerin mekanik özelliklerinin tahmininde deneysel testleri destekleyen güçlü bir mühendislik aracı olabileceğini ortaya koymaktadır.

Keywords

• Ahşap Kiriş
• Eğilme Dayanımı
• Ahşap Malzeme
• Tutkal
• Makine Öğrenme

Prediction of Bending Strength (MOR) Values of Laminated Timber Beams Using Machine Learning Models

Abstract

This study aims to determine the bending strength (MOR) values of glued laminated timber (glulam) beams with different geometries and production parameters through three-point (3P) and four-point (4P) bending tests and to predict these values using machine learning-based models. The experimental beams were produced in four different groups: straight, curved, plywood-reinforced, and unreinforced, using PVAc-D3 (Polyvinyl Acetate) adhesive-bonded Scots pine (Pinus sylvestris L.) lamellas. The experimental data obtained were analysed using five different regression algorithms (SVR, Lasso, Ridge, Huber, and XGBoost), and the prediction performance of each model was evaluated using 10-fold cross-validation. According to the analysis results, the SVR model showed the most successful performance in the three-point bending tests with quite high accuracy and low error values such as R² = 0.9920 and RMSE = 0.4587 MPa. A similar trend was observed in the four-point tests, with the SVR model maintaining its superior performance with R² = 0.9782. These findings demonstrate that machine learning-based approaches can be a powerful engineering tool to support experimental testing in predicting the mechanical properties of glulam beams.

Keywords

• Wooden Beam
• Bending Strength
• Machine Learning
• Wooden Material
• Glue

References

1. K. Sikora, D. McPolin, and A. Harte, “Mechanical performance of Glulam and CLT beams,” Construction and Building Materials, vol. 102, pp. 110–120, 2016.
2. M. H. Ramage, H. Burridge, M. Busse-Wicher, et al., “The wood from the trees: The use of timber in construction,” Renewable and Sustainable Energy Reviews, vol. 68, pp. 333–359, 2017.
3. J. Abed, S. Rayburg, J. Rodwell, and M. Neave, "A review of the performance and benefits of mass timber as an alternative to concrete and steel for improving the sustainability of structures," Sustainability, vol. 14, no. 9, p. 5570, 2022, doi: 10.3390/su14095570.
4. O. Perçin, G. Özbay, and M. Ordu, “Farklı tutkallarla lamine edilmiş ahşap malzemelerin mekaniksel özelliklerinin incelenmesi,” Journal of Science and Technology of Dumlupınar University, no. 19, pp. 109–120, 2009.
5. TS EN 408 + A1 Ahşap yapılar - Yapısal ahşap ve lamine ahşap - Fiziksel ve mekanik özelliklerin tayini. Ankara: Türk Standardları Enstitüsü, 2014.
6. TS ISO 13061-3 Odunun fiziksel ve mekanik özelliklerinin tayini - Bölüm 3: Statik eğilme dayanımının tayini*. Ankara: Türk Standardları Enstitüsü, 2021.
7. A. Derkowski, M. Kuliński, A. Trociński, J. Kawalerczyk, and R. Mirski, “Mechanical characterization of glued laminated beams containing selected wood species in the tension zone,” Materials, vol. 15, no. 18, p. 6380, 2022.
8. D. Dziurka, A. Derkowski, M. Wieruszewski, M. Kuliński, and R. Mirski, “GL beams reinforced with plywood in the outer layer,” Materials, vol. 15, no. 11, 3976, 2022.

9. R. Bülbül, M. Kaya, H. İmirzi, and N. Döngel, "Farklı Odun ve Tutkal Türleri Kullanılarak Üretilen Lamine Kerestelerin (Glulam) Bazı Fiziksel ve Mekanik Özellikleri," Bartın Orman Fak. Derg., vol. 26, no. 2, pp. 58-68, 2024, doi: 10.24011/barofd.1403573.
10. Y. Liu, H. Zhang, and M. Chen, “Machine learning in materials science: Recent progress and emerging applications,” Materials Today, vol. 41, pp. 56–69, 2020.
11. S. Elhishi, A. M. Elashry, and S. El-Metwally, “Unboxing machine learning models for concrete strength prediction using XAI,” Scientific Reports, vol. 13, no. 1, 19892, 2023.
12. S. S. Sorour, C. A. Saleh, and M. Shazly, “A review on machine learning implementation for predicting and optimizing the mechanical behaviour of laminated fiber-reinforced polymer composites,” Heliyon, vol. 10, no. 13, 2024.
13. H. Hu, Q. Wei, T. Wang, Q. Ma, P. Jin, S. Pan, et al., “Experimental and numerical investigation integrated with machine learning (ML) for the prediction strategy of DP590/CFRP composite laminates,” Polymers, vol. 16, no. 11, 1589, 2024.
14. M. Yossef, M. Noureldin, and A. Alqabbany, “Explainable artificial intelligence framework for FRP composites design,” Composite Structures, vol. 341, 118190, 2024.
15. P. Xu, X. Ji, M. Li, and W. Lu, “Small data machine learning in materials science,” npj Computational Materials, vol. 9, no. 1, 42, 2023.
16. F. Oviedo, J. L. Ferres, T. Buonassisi, and K. T. Butler, “Interpretable and explainable machine learning for materials science and chemistry,” Accounts of Materials Research, vol. 3, no. 6, pp. 597–607, 2022.
17. S. Chen, R. Shiina, K. Nakai, T. Awano, A. Yoshinaga, and J. Sugiyama, "Potential of machine learning approaches for predicting mechanical properties of spruce wood in the transverse direction," J. Wood Sci., vol. 69, no. 1, p. 22, 2023, doi: 10.1186/s10086-023-02094-1.
18. A. R. Haftkhani, F. Abdoli, A. Sepehr, and B. Mohebby, “Regression and ANN models for predicting MOR and MOE of heat-treated fir wood,” Journal of Building Engineering, vol. 42, 102788, 2021.
19. S. Bardak, T. Bardak, H. Peker, E. Sözen, and Y. Cabuk, “Predicting effects of selected impregnation processes on the observed bending strength of wood, with use of data mining models,” Bioresources, vol. 16, no. 3, pp. 4891–4908, 2021.
20. M. Nazerian, F. Naderi, and A. N. Papadopoulos, “Application of the artificial neural network to predict the bending strength of the engineered laminated wood produced using the hydrolyzed soy protein-melamine urea formaldehyde copolymer adhesive,” Journal of Composites Science, vol. 7, no. 5, 206, 2023.
21. M. Lukovic, L. Ciernik, G. Müller, D. Kluser, T. Pham, I. Burgert, and M. Schubert, “Probing the complexity of wood with computer vision: From pixels to properties,” Journal of the Royal Society Interface, vol. 21, no. 213, 20230492, 2024.
22. J. Ma, Z. Kuang, Y. Fang, and J. Huang, “A multi-input residual network for non-destructive prediction of wood mechanical properties,” Forests, vol. 16, no. 2, 355, 2025.
23. J. Huang and Z. Kuang, “Mechanical property prediction of wood using a backpropagation neural network optimized by adaptive fractional-order particle swarm algorithm,” Forests, vol. 16, no. 8, 1223, 2025.
24. S. Kayakıran and E. Kishalı, “Dünden bugüne tutkallı tabakalı ahşap yapı elemanların incelenmesi ve yapılarda taşıyıcı olarak kullanılması üzerine öneriler,” Journal of Architectural Sciences and Applications, vol. 4, no. 1, pp. 34–50, 2019.
25. R. Becer en Öztürk and N. Arioğlu, “Türk sarıçamından lamine ahşap kirişlerin mekanik özellikleri,” İTÜDERGİSİ/a, vol. 5, no. 2, 2010.
26. A. Kasal, H. Efe, and T. Dizel, “Masif ve lamine edilmiş ağaç malzemelerde eğilme direnci ve elastikiyet modülünün belirlenmesi,” Politeknik Dergisi, vol. 13, no. 3, pp. 183–190, 2010.
27. Y. Bozkurt and N. Erdin, Odun Anatomisi, İstanbul Üniversitesi Orman Fakültesi Yayın No. 4263/466, İstanbul, 2000.
28. R. Kurt, “Ahşap yapıştırma ve yapıştırıcılar,” in Ahşap Malzeme Bilgisi, Y. Ünsal, Ed., Gazi Kitabevi, Ankara, 2018, pp. 215–240.
29. N. Döngel, Bazı ağaç türlerinden farklı tutkallarla üretilen lamine elemanların eğilme direnci üzerine bir araştırma, Yüksek Lisans Tezi, Karadeniz Teknik Üniversitesi, Fen Bilimleri Enstitüsü, Trabzon, 1999.
30. DMC Endüstri A.Ş., PVAc – D3/D4 ahşap tutkalı ürün teknik bilgi föyü (TDS), DMC Endüstri A.Ş., 2024.
31. A. Zalcmanis, K. Abols, K. Zudrags, and A. Krasnikovs, “Birch plywood sample cyclic bending property investigation and analysis,” in Engineering for Rural Development. Proceedings of the International Scientific Conference (Latvia), no. 19/2020.
32. F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, et al., “Scikit-learn: Machine learning in Python,” Journal of Machine Learning Research, vol. 12, pp. 2825–2830, 2011.
33. T. Chen and C. Guestrin, “XGBoost: A scalable tree boosting system,” in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 785–794, 2016.
34. A. J. Smola and B. Schölkopf, “A tutorial on support vector regression,” Statistics and Computing, vol. 14, no. 3, pp. 199–222, 2004.
35. A. E. Hoerl and R. W. Kennard, “Ridge regression: Biased estimation for nonorthogonal problems,” Technometrics, vol. 12, no. 1, pp. 55–67, 1970.
36. R. Tibshirani, “Regression shrinkage and selection via the LASSO,” Journal of the Royal Statistical Society: Series B (Statistical Methodology), vol. 58, no. 1, pp. 267–288, 1996.
37. P. J. Huber, “Robust estimation of a location parameter,” in Breakthroughs in Statistics: Methodology and Distribution, New York, NY: Springer, pp. 492–518, 1992.
38. R. Kohavi, “A study of cross-validation and bootstrap for accuracy estimation and model selection,” in Proceedings of the 14th International Joint Conference on Artificial Intelligence (IJCAI), vol. 14, no. 2, pp. 1137–1145, 1995.
39. R. J. Hyndman and A. B. Koehler, “Another look at measures of forecast accuracy,” International Journal of Forecasting, vol. 22, no. 4, pp. 679–688, 2006.
40. T. Chai and R. R. Draxler, “Root mean square error (RMSE) or mean absolute error (MAE)? – Arguments against avoiding RMSE in the literature,” Geoscientific Model Development, vol. 7, no. 3, pp. 1247–1250, 2014.
41. C. J. Willmott and K. Matsuura, “Advantages of the mean absolute error (MAE) over the root mean square error (RMSE),” Climate Research, vol. 30, no. 1, pp. 79–82, 2005. J. S. Armstrong, Long-Range Forecasting: From Crystal Ball to Computer, 2nd ed., Wiley, New York, 1985.
42. S. Kim and H. Kim, “A new metric of absolute percentage error for intermittent demand forecasts,” International Journal of Forecasting, vol. 32, no. 3, pp. 669–679, 2016.
43. N. R. Draper and H. Smith, Applied Regression Analysis, vol. 326, John Wiley & Sons, 1998.